Sie sind auf Seite 1von 255



JAMES R. LIBBY, President

NORMAN D. PERKINS, Vice President

An old man going a lone highway

Came at the evening cold and gray
To a chasm vast and deep and wide.
The old man crossed in the twilight dim;
The sullen stream had no fears for him.
But he turned when safe on the other side
And built a bridge to span the tide.

“Old man,” said a fellow pilgrim near,

“You are wasting your time with building here.
You never again will pass this way-
Your journey will end with the closing day.
You have crossed the chasm deep and wide,
Why build you this bridge at eventide?”

The builder lifted his old gray head.

“Good friend, in the way that I’ve come,” he said,
“There followeth after me today
A youth whose feet must pass this way.
This stream which has been as naught to me
To the fair-haired youth might a pitfall be.
He, too, must cross in the twilight dim.
Good friend, I’m building the bridge for him.”

Will Allen Dromgoole


This book has been written with the intention of describing the fundamen-
tal structural behavior of the most commonly used prestressed concrete
bridges. The authors believe the contents of this book will be especially
useful to engineers having little or no previous experience in the design of
prestressed concrete bridges as well as those whose practice includes an
occasional bridge design.
The first chapter is devoted to basic information and serves as a founda-
tion for subsequent chapters.
Chapter 2 is devoted to girder bridges. The authors elected to use this
name over “stringer bridge” in view of the fact that the term “stringer” is
not applied to beams of reinforced or prestressed concrete in the Standard
Specifications for Highway Bridges which is published by the American
Association of State Highway and Transportation Officials. This form of
concrete bridge has been the type most commonly used in the United
States. Its use has been widespread and is expected to continue. Methods
of analysis for girder bridges which have been in use in Europe for a number ’
of years are presented in this chapter. These methods have not been
commonly used in this country because they are not usually taught in our
universities. In addition, they are not included in the bridge design criteria
normally used in this country. The significant effect of well designed
transverse beams or diaphragms on the distribution of live loads to the
individual girders is emphasized.
Box-girder bridges are treated in Chapter 3. This important form of
cast-in-place construction has been widely used in the western United
States. Its use in other parts of the country is increasing and is expected to
reach very significant levels in the next few years. The importance of the
torsional stiffness of the box-girder cross section is explained as is its effect
on the distribution of flexural stresses due to live loads.
A relatively new form of concrete bridges has been treated in Chapter 4.
It has been referred to as a segmental box-girder or a segmental bridge in
this book. Design considerations and construction techniques unique to
this mode of bridge construction are treated in detail. This chapter contains
information that should be of value to experienced bridge designers as well
as to those without extensive experience.
The additional design considerations of Chapter 5 and the construction
considerations of Chapter 6 have been included as a means of calling the
reader’s attention to a number of factors requiring consideration in for-
mulating a complete bridge design. Some of the subjects included may not
be new or may be apparent to some of the readers. Others will find these
chapters convenient sources of reference from time-to-time.
The authors wish to acknowledge the technical information and photo-
graphs that have been provided by the French engineering firm, Europe
Etudes. In particular, the contributions of Jean Muller and Gerard
Sauvageot are acknowledged with sincere thanks. The authors also wish to
thank the publishing firm of Springer-Verlag for permission to publish the
influence surface charts reproduced in this book as Figures 1.2 through 1.5
and Jacob Dekema of the California Department of Transportation for the
excellent photographs of bridges designed and constructed under supervi-
sion of the Department.
San Diego, California James R. Libby
September, 1975 Norman D. Perkins


Preface V

1.1 Scope of book 1
1.2 Design criteria 2
1.3 Design loads 3
1.4 Design methods 5
1.5 Allowable stresses 13
1.6 Bridge types considered 15
1.7 Span length vs. bridge type 28

2.1 Introduction 29
2.2 Girder design 42
2.3 Intermediate diaphragms 44
2.4 Decks for girder bridges 46
2.5 Continuity 48
2.6 Overhanging beams 49
2.7 Construction details 50

3.1 Introduction 57
3.2 Flexural analysis 59
3.3 Longitudinal flexural design 66
3.4 Decks for box-girder bridges 67
3.5 Shear distribution 69
3.6 Construction details 73 ’

4.1 Introduction 83
4.2 Longitudinal flexural analysis 85
4.3 Creep redistribution of moments 96
4.4 Transverse flexure 100
4.5 Proportioning the superstructure 119
4.6 Proportioning the segment 126
4.7 Intermediate hinges 134
4.8 Support details 135
4.9 Construction details 141


5.1 Introduction 145
5.2 Design for shear 145
5.3 Horizontally curved bridges 150
5.4 Strength analysis 151
5.5 Elastomeric bearing pads 154
5.6 Substructure considerations 161
5.7 Seismic forces 173

6.1 Introduction 175
6.2 Falsework , 176
6.3 Formwork 179
6.4 Concete finishes 187
6.5 Erection of precast girders 188
6.6 Erection of precast segments 188
6.7 Camber control 195
6.8 Quantity of prestressing material 203

A. Long-term Deformation of Concrete 205
B. Standard Shapes of Precast Beams 213
C. Analysis of Statically Indeterminate Structures With the
Method of Support Constants 217
D. Thermal Stresses In Concrete Bridge Superstructures 241

References 247

Index 251

1 ‘I Moduction

1.1 Scope of Book.

This book has been written with the principal purpose of describing the
design methods that are applicable to the various major types of prestres-
sed concrete highway bridge superstructures currently in use in the United
States. Secondary purposes have been to describe the advantages and
disadvantages of the various bridge types and to briefly discuss the con-
struction methods used with the different types.
Reinforced concrete bridge superstructures are not considered. The
basic principles of elastic design which are discussed in this book are,
however, equally applicable to reinforced concrete and prestressed con-
Bridge substructure design is considered only as it affects the design of
the bridge superstructure or the bridge as a whole.
The fundamental principles of reinforced concrete and prestressed con-
crete structural design are not presented in this book. It is presumed the
reader is competent in the design of these forms of concrete construction
(Ref. 1,2).* In addition, it is presumed the reader is familiar with the

*I, 2, etc. refer to references listed in the back of this book.


strength, elastic, creep and shrinkage properties of portland cement con-

crete as well as with the properties ofordinary reinforcing steel (Ref. 3) and
the steels commonly used in the United States in prestressed concrete
construction (Ref. 4,5). Finally, it is presumed the reader is familiar with
the fundamental principles of structural analysis.
Not all types of prestressed concrete bridge superstructures used or
proposed for use in the United States have been included in this book.
Bridges which employ -precast members used primarily in building con-
struction, such as double-tee beams, single-tee beams, hollow-core slabs
and solid precast slabs, are discussed briefly, but are not treated in detail.
No attempt has been made to include a discussion of unique bridges
utilizing specially fabricated precast sections peculiar to the specific bridge
even though the bridge might be considered to be a major structrue. Many
ofthe fundamental principles discussed in this book are, however, equally
applicable to structures of these types.
Specific cost data have not been included in the discussions of the
various types of bridges. Construction costs vary with the constantly
changing economy of the nation. The result is that specific cost data are
normally only accurate for a short period of time. Relative construction
costs vary throughout the country and hence escape anything but vague

1.2 Design Criteria.

The most widely used criteria for the design and construction of highway
bridges in North America are contained in the “Standard Specifications for
Highway Bridges” (Ref. 6) published by the American Association of
State Highway and Transportation Officials.* These criteria, which are
referred to subsequently in this book as the “AASHTO Specification”, or
simply as “AASHTO”, are used as the basic criteria for design except
where otherwise stated.
/ The design criteria pertaining to reinforced and prestressed concrete
contained in the AASHTO Specification are based to some degree upon
the American Concrete Institute publication “Building Code Require-
ments for Reinforced Concrete” (AC1 318) (Ref. 7). This publication is
referred to subsequently as AC1 318. In some instances specific references
to this publication are made in the AASHTO Specification. This AC1
publication, which is under constant review and frequent revision, reflects

*Previous to the year 1974, this organization was known as the American Association of State
Highway Officials.

the best contemporary thinking relative to the design of concrete struc-

The designer of prestressed concrete bridges should be familiar with the
provisions of the latest editions (with interim modifications) of both the
AASHTO Specification and AC1 318. His design should incorporate the
provisions of these publications which will result in a safe structure that
behaves in a predictable manner.
AC1 Committee 443, Concrete Bridge Design, has published two por-
tions of what eventually will become a complete recommended practice for
the design of concrete bridges (Ref. 8,9). These publications are highly
recommended to all who are interested in the design of concrete bridges.

1.3 Design Loads.

Like other structures, bridges must be designed for the dead and live loads
to which they are subjected.
The dead loads consist of the self-weight of the basic structural section
itself as well as superimposed dead loads such as bridge railings, sidewalks,
non-structural wearing surfaces, and utilities which the structure must
support. Dead loads can generally be estimated with a high degree of
accuracy during the design, accurately controlled during the construction
and are normally considered to be permanent loads. Due to their more or
less permanent nature, loads resulting from concrete volume changes are
sometimes categorized as dead loads.
Live loads are those due to the effect of external causes and are generally
transient in nature. Live loads include those resulting from vehicles and
pedestrians which pass over the bridge as well as the forces resulting from
wind, earthquake and temperature variation. Other live loads are secon-
dary in nature and result from impact forces. Vertical impact forces are
created by the vehicles using the structure. Horizonal impact forces result
from braking and turning of these vehicles. The live loads that will be
imposed upon a structure cannot generally be estimated with the same
precision as can the dead loads. In addition, the designer often has little if
/ any control over these loads once the structure is put into service.
The minimum live loads for which bridge structures must be designed are
generally specified by design criteria such as the AASHTO Specification.
/ Considerable differences exist in the live load design criteria used through-
out the world. Much has been written on this as well as on the fact that the
criteria used in the United States may be unrealistically low and may not be
representative of the actual loads to which our bridges are exposed (Ref.
10,ll). From these discussions the bridge designer should keep two facts in
4 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

mind. These are: (1) the live load requirements specified by the AASHTO
Standard Specification are among the lightest loadings used in the world;
and (2) these live load requirements may be lower than the maximum loads
one might expect on a highway bridge in the United States.
It may very well be that other requirements of the AASHTO Specifica-
tions compensate to some degree for the relatively light design live loads
specified therein. Some engineers feel the day has come for the AASHTO
Standards to be materially revised with a view toward specifying’ more
realistic truck loadings as well as encouraging more sophisticated methods
of bridge design and analysis. Zf the design live loads of the AASHTO
StandardSpecification are too low, they should be increasedso that elastic
analyses will yield reasonable agreement with what is actually occurring in
real bridges. One should not rely upon the conservatism of empirical
coeficients to compensate for inadequate load criteria. This is especially
true when strength rather than service load design methods are used.
The design loads that must be considered in the design of reinforced
concrete and prestressed concrete are identical except for those caused by
volume changes. The effect of concrete shrinkage is less in the case of
reinforced concrete than in the case of prestressed concrete. This is due to
the fact that non-prestressed reinforcing steel tends to resist concrete
shrinkage strains and, in reinforced concrete members, promotes the for-
mation of fine cracks. The fine cracks relieve the shrinkage stresses in the
concrete as well as the need for the member to shorten. The important
effect of concrete creep on reinforced concrete members is the time-
dependent effect on deflection. In prestressed concrete the cracking
mechanism related to shrinkage does not take place and provision must be
made for the total shrinkage strain which may occur and cause undesirable
effects. In prestressed concrete creep and shrinkage both affect deflection.
This must be considered in the design. Shortening due to creep and shrin-
kage can be significant in prestressed concrete structures and must be
taken into account if good results are to be obtained.
Although there are considerable data in the literature relative to creep
and shrinkage of concrete, there is no accepted U.S. recommended prac-
tice for estimating the magnitude of the creep and shrinkage strains the
designer should accommodate in his design. Methods have been proposed
in the literature (Ref. 12,13) but these have not achieved the status of a
standard or recommended practice. For the benefit of the reader, the
methods used for predicting concrete shrinkage and creep in the French
Code (Ref. 14) are included as Appendix A of this book.
Due to the complexity of the live load criteria given in the AASHTO
Specification, these provisions will not be repeated in this book. Most
bridges are designed for the AASHTO HS20-44 live load. Live loads of

lower magnitude than HS20-44 are provided in the AASHTO Specifica-

tion. The smaller live loads were originally intended for use on secondary
roads but not on primary highways. Because there is virtually no practical
way of insuring that the largest trucks will not be used on secondary roads,
many jurisdictions use the HS20-44 loading in all bridge design.
The AASHTO Specification stipulates that a truck or lane loading shall
be assumed to occupy a width of ten feet. Each IO-foot wide truck or lane
load is to be positioned in a design traffic lane which is twelve feet wide. It
is further stipulated that the number of design traffic lanes shall be two for
bridges having roadway widths between curbs of from 20 to 24 feet. For
roadway widths over 24 feet, each design traffic lane is assumed to occupy
a width of 12 feet. The twelve-foot width traffic lanes are to be positioned in
such a manner as to produce the maximum stress in the member under
’ consideration. For bridges which are designed for three or more design
traffic lanes, Section 1.2.9 of the AASHTO Specification provides load
intensity reduction factors which are intended to account for the improba-
bility of all lanes being frequently loaded simultaneously. The location of
the specific live load-related requirements of the AASHTO Specification
are summarized in Table 1.

TABLE l-Factors in the AASHTO Specification related to live load design criteria.
H Truck & Lane loadings, dimensions & loads Para. 1.258
HS Loadings, dimensions & loads Para. 1.2.5C
Traffic Lanes. number and width See Interim 1, 1974 Interim
Specification Bridges of AASHTO.
Prohibition of use of fractional truck
and loadings. Para. 1.2.8A and Interim 1, 1974 Interim
Specifications Bridges of AASHTO.
Use of truck versus lane loadings Para. 1.2.88
Continuous spans-modification to lane
loadings Para. 1.2.8C
Loading for maximum stress Para. 1.2.8D
Reduction in load intensity, multi-lane
structures Sect. 1.2.9
Sidewalk, Curb and Railing Loading Sect. 1.2.11
Impact Loading Sect. 1.2.12

1.4 Design Methods

Empirical coefficients are included in the AASHTO Specification for

determining the live load moments for which concrete deck slabs are to be
designed as well as for determining the distribution of live loads to the \
girders which support the concrete slabs. The use of the empirical coeffi-
6 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

cients is not mandatory but they are given for use when more sophisticated
methods of analysis are not used.
The live load distribution factors which are contained in Section 1.3.1 of
the AASHTO Specification, are based upon the assumption a bridge can
be divided into several longitudinal beams for the purpose of design and
analysis. A bridge is, of course, a three dimensional structure and should
be designed with this being taken into account. Approximate elastic
methods of analysis, which consider the superstructure as a whole, are
presented in Chapters 2,3 and 4 for use in the design of bridge superstruc-
tures which are narrow with respect to their span as well as relatively deep
with respect to their width. The approximate methods are applicable to
most conditions encountered in practice.
Sophisticated methods of analyzing bridge structures including the


10 15 20 25

Span (Ft.)
Fig. 1 .l Comparison of the design moment required by the 1957 and 1961 AASHTO Specifi-
cations for concrete bridge decks with their main reinforcement perpendicular to the
direction of traffic, together with elastic analyses for two specific conditiont of
loading. Simple spans, no impact.

folded plate method, the finite segment method and the finite element
method are described in the literature (Ref. 15). These methods result in
higher precision in the determination of the stresses and deflections than
can be obtained by use of the familiar flexural theory. These methods have
a place in structural research and in the design of special structures but
their use is not needed nor considered to be practical in normal design
work. The slightly greater accuracy in determining stresses with these
methods in bridges of normal proportions is not significant when one
considers the differences between the loads used in design and the actual
loads to which a structure can be subjected. The cost of employing the
more sophisticated methods as a design procedure is prohibitive in most
For many years the design of bridge decks has been done using empirical
relationships contained in the AASHTO Specification. Relationships
which are based upon the work of H. M. Westergaard (Ref. 16), are given
for slabs which have their main span perpendicular to the direction of
traffic as well as for slabs which have their main spans parallel to the
direction of traffic.
A major revision of the empirical relationships for the design of bridge
slabs occurred in the interim between the 1957 and 1961 AASHTO Speciti-
cations. A comparison of the design requirements for simply supported
slabs having their main reinforcement perpendicular to the direction of
traffic according to the 1957 and 1961 AASHTO Specifications for HS20-
44 live loads without impact is given in Fig. 1.1.
Two basic design deficiencies exist with these empirical relationships.
The first of these is the lack of provisions to account for the differences
between the distributions of positive and negative moments in members
having constant and variable depth. The second is the moment continuity
coefficient of 0.80 which is specified for decks which are continuous over
three or more supports regardless of the elastic restraints which are pro-
vided by the various members which are connected at the supports of the
The empirical relationships ofthe AASHTO Specifications have proved
to be satisfactory for the decks of bridges which are supported by torsion-
ally flexible stringers that may or may not be connected together with
flexurally stiff transverse diaphragms. Hence, these relationships can be
considered to be conservative for all structural schemes. The relationships
may, however, be overly conservative with respect to the moments in
decks that are supported by torsionally flexible stringers connected with
flexibly stiff intermediate diaphragms or which are supported by torsion-
ally stiff systems. Additionally, the empirical relationships do not alert the ,
designer to the importance of considering the live load deck moments
. -\ I 1. :._ ‘.; ‘.
/------_ ‘---~---’ e____-___ \ \ ._,.

Fig. 1.2 Influence surface for the midspan moment in the xdirection of a plate strip of constant
depth with two restrained edges and Poisson’s ratio = 0 (8 T times the actual values
shown). (Courtesy Springer-Verlag).

induced in the webs of flexurally stiff supports. Hence, in this respect they
are unconservative.
In view of the above, elastic design methods are recommended for decks
of most concrete bridges.
Charts of influence surfaces, which can be thought of as being similar to
two-dimensional influence lines, are available for the determination of
moments, shears and deflections for slabs having a variety of dimensions
and boundary conditions (Ref. 17,18). Examples ofthe charts* are given in
Figs. 1.2 through 1 S. The charts are used by plotting the “footprints” of
the applied wheel loads, adjusted to the proper scale, on the charts and
computing the volumes defined by the area of the “footprints” and the
ordinates of the chart. The sum of the products of the volumes and their
respective loads is equal to the moment, the shear or the deflection coefft-
cient for which the chart has been prepared. The charts are prepared with
the assumption that Poisson’s ratio, for the material of which the slab is
composed, is equal to zero. Hence, a correction factor must be applied to
the computation to correct for this assumption. The instructions which are
included with the charts explain how this correction should be made. The
charts are based upon an elastic analysis. The moments or shears com-
puted by use of,the charts are expressed per unit of length (i.e. Kip-feet per
foot or kips per foot for moment and shear respectively) at the location in
the slab for which the chart was prepared.
The charts presented by Pucher are for slabs of constant depth only
while those prepared by Homberg include charts for slabs of constant as
well as variable depth.
The use of influence charts permits the designer to take the effects of
variable slab thickness into account. In addition, because the charts are
based upon rational elastic analysis, they permit the designer to analyze the
effects ofjoint restraint. This is accomplished by determining the fixed-end
moments for the critical conditions of loading and distributing them to the
supporting members in accordance with normal elastic design procedures.
As was stated above, the empirical relationships for the design of bridge
decks which are contained in the AASHTO Specifications do not give the
designer a basis for taking either of these factors into account.
The design span to be used in the design of prismatic solid slabs which
are constructed monolithically with their supports is generally taken as the
clear distance between supports. This assumption is limited to spans of 10
feet or less by the AC1 Building Code Requirements for Reinforced Con-
crete (Ref. 19) but not by the AASHTO Specification (Ref. 20). For spans
in excess of 10 feet, according to the AC1 Requirements, one should base

*Courtesy of Springer-Verlag.
Fig. 1.3 Influence surface for the moment at the support in the x direction for a cantilevered
plate strip of constant depth and Poisson’s ratio = 0 (8 n times the actual values
shown). (Courtesy Springer-Verlag).

P ~


Fig. 1.4 Influence surface for the moment at the support in the x direction for a plate strip of
variable depth (parabolic) with two restrained edges and Poisson’s ratio = 0.
(Courtesy Springer-Verlag).

Fig. 1.5 Influence surface for the support moment in the x direction for a cantilevered plate
strip of variable depth (constant variation from d to 2d) and Poisson’s ratio = 0.
(Courtesy Springer-Verlag).

the determination of moments upon center-to-center distances between

joints but use the moments computed at the faces of supports in designing
the slab for strength.
Little guidance is to be found in the usual structural design criteria
employed in the United States relative to the design span to be used when
haunched members are employed. The definition of the design span per-
mitted by the French Code is shown in Fig. 1.6 for various conditions of
haunches (Ref. 21). The French Code limits these definitions of design
spans to spans of 6 meters (19.7 feet) or less and to slabs which are
supported on their entire (or nearly so) perimeter and which are subjected
to large transient concentrated loads. The complete provisions of the
French Code relative to the design of slabs is recommended for reading but
is not reproduced here.

1.5 Allowable Stresses

The allowable stress criteria that are followed in a structural design obvi-
ously have a major influence on the results. If, for example, no flexural
tensile stresses are to be permitted in a prestressed concrete flexural
member under full dead, live and impact loading, significantly more pre-
stressing steel may be required in the member than would be required if the
allowable stresses permitted by Section 1.6 of the AASHTO Standards
were followed.
In spite of the fact that Section 1.6.6 of the AASHTO Specifications
permits tensile stresses in prestressed concrete members, some engineers
feel tensile stresses should not be permitted in certain instances. One of
these is in the design of bridge decks. It has been argued that wheel loads in
some instances, whether legally or illegally, exceed the design loads
specified by the AASHTO Specification. Iftensile stresses were permitted
when designing for the AASHTO wheel loads, the tensile strength of the
concrete could be exceeded if the slabs were subjected to wheel loads
greater than the AASHTO wheel loads. Some engineers believe that
segmentally constructed bridges, which are treated in detail in Chapter 4,
should be designed as Class I structures. * This opinion is based upon the
belief that the many joints in segmental bridges makes them more suscepti-
ble to deterioration than is the case for monolithic structures and hence
more conservative stresses are indicated. In addition, the redistribution of
*Current European practice is to categorize structures into three classes. The distinguishing
factor is the allowable tensile stress and hence degree of cracking, which is permitted.
Tensile stresses are not permitted in Class I structures. Tensile stresses as high as the tensile
strength of the concrete are permitted in Class 11 structures. Tensile stresses which exceed ’
tensil strength of the concrete are permitted in Class II structures (Ref. 22).


F i g . 1.6 Method of determining design span for haunched slats according to the French code
for prestressed concrete.

moments which occurs due to concrete creep in segmental bridges that are
erected in cantilever is sometimes estimated by approximate calculation
rather than being determined with precision and hence conservatism in
areas of positive moment seems appropriate. Finally, the existence of a
temperature differential between the deck and girders of a girder bridge or
between the top slab and the remaining portions of a box-girder bridge
results in the creation of moments and hence flexural stresses in a struc-
ture. The stresses due to differential temperature in a simple-span structure
may be of relatively nominal magnitude and are directly dependent upon
the magnitude of the temperature differential. In the case of continuous
structures, it can be shown that nominal temperature differentials can
result in flexural tensile stresses as great as 500 psi and can result in
significant temperature induced variations in the reactions at the beam

supports (Ref. 23,24). Stresses as great as these should be considered in

service load analyses of bridges in which flexural tensile stresses are
permitted under live and impact loads. Methods of evaluating the effect of
temperature variations within a section are given in Appendix D.
One should bear in mind that the provisions of the AASHTO Specifica-
tions, like most design criteria engineers are required to observe in their
work, constitutes a minimum criteria. An engineer, using his own judge-
ment and experience as a basis, may wish to follow more conservative
criteria in his work.

1.6 Bridge Types Considered

Three types of bridges are considered in detail in this book. The design
considerations unique to each of the three methods of construction are
treated in Chapters 2, 3 and 4. Construction considerations for each of the
three types of bridges are treated in Chapter 6. The bridge types considered
in detail are referred to as girder bridges, box-girder bridges and segmental

7 k
Fig. 1.7
Typical cross section of a T-beam bridge.

28’-0” \
Fig. 1.8 Typical cross section of a precast l-beam bridge.


1 ?z’: Q HOLE IN WEB



Fig. 1.9 Typical end diaphragm details for an l-beam bridge.

bridges. Other types of prestressed concrete bridges are not discussed in

detail either because they will behave structurally in a manner that is
similar to one of the “basic” three bridge types or because their design is
normally accomplished with sufficient accuracy using the empirical rela-
tionships included in the AASHTO Specification.
Girder bridges are bridges which incorporate two or more longitudinal
beams togetherwith a deck slab spanning transversely over or between the
top flanges of the girders. The deck slab is normally connected to the top
flanges of the girders in such a manner that it acts compositely with them in
resisting a portion, if not all, of the longitudinal flexural stresses. It is not
essential that the deck slab acts compositely with the girders. Girder
bridges are frequently constructed as simple spans. They are also fre-
quently constructed continuous over two or more spans. Bridges of this
type have found wide use in North America.
Typical cross sections of two types of girder bridges are shown in Figs.
1.7 and 1.8. The first of these is typical of cast-in-place construction and is
often referred to as a “T-beam” structure. (Rectangular precast beams can
be used, in combination with a cast-in-place deck slab, to form a bridge of
this type.) The longitudinal beams are normally considered to have effec-
tive cross sections which are T-shaped in the analysis for longitudinal
flexure. The second cross section (Fig. 1.8) is typical of bridges formed of
precast I-shaped or T-shaped beams together with a cast-in-place deck.
An important characteristic of the beams used in bridges of this type is
their relatively small torsional stiffness. When sophisticated methods of

Ll” a x 16” BOLT IN


Fig. 1 .lO Typical intermediate diaphragm details for an l-beam bridge.

analysis are used to analyze girder bridges, the torsional stiffness of the
beams is normally neglected. This is treated in greater detail in Chapter 2.
Transverse beams, which are called end diaphragms, are normally pro-
vided at each support of girder bridges. The end diaphragms connect the
longitudinal girders to each other as well as to the deck and provide an
efficient means of transferring lateral loads, acting upon the superstruc-
ture, to the substructure. The end diaphragms also prevent movement of
the ends of the beams with respect to each other. Typical end diaphragm
details are shown in Fig. 1.9.
Diaphragms are usually provided between the girders at one or more
locations between supports. These diaphragms are termed intermediate

i 28’-0”
I- I

Fig. 1 .l 1 Typical cross section of a box girder bridge.

1 8 )H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

diaphragms. Intermediate diaphragms are frequently not used in bridges

having spans of 40 feet or less. Commonly used details for intermediate
diaphragms in bridges utilizing I-shaped beams are given in Fig. 1.10. (See
Section 2.3)
A typical cross section for a box-girder bridge is shown in Fig. 1.11. A
considerable number of bridges of this type has been constructed in the
United States. This is especially true for the Western United States.
Box-girder bridges are cast-in-place on falsework. Superstructures of
two or more spans are normally made continuous over the interior sup-
ports. They are frequently constructed with fixed connections between the
superstructure and the abutments, piers or bents and hence form a frame.
The use of frames has been considered important in regions where
earthquakes might be expected.
Construction joints are normally provided in box-girder bridges near the
junction between the webs and the upper slab as shown in Fig. 1.11. The
bottom slab and web stems are usually constructed at one time. After the
formwork for the interior webs has been removed, forms for the upper deck
are installed and the upper deck is constructed. The forms used for the
upper deck at interior cells are generally left in place.
End diaphragms are used with box-girder bridges as they are for girder
bridges and for the same reasons. Although intermediate diaphragms have
been used on many box-girder bridges, they serve little if any useful
function (except in structures having significant horizontal curvature) due
to the great torsional stiffness as well as the transverse flexural stiffness of
the box-girder section. The fact that intermediate diaphragms are not
needed in box-girder bridges is currently recognized by many bridge
engineers and their use is expected to diminish rapidly in the future (Ref.
The torsional stiffness of the box-girder bridge superstructure as well as
the transverse flexural stiffness are important structural characteristics of
this mode of bridge construction. This is considered in greater detail in
Chapter 3.
Specific forms of box-girder bridges are referred to in this book as
“segmental box-girder bridges” or as “segmental bridges”. Some may feel
this distinction is not necessary or justified. The authors believe the dis-
tinction is not only justified but necessary at this time if the full potential of
this form of bridge construction is to be realized. The use of a special name
for this mode of construction will facilitate calling the attention of the
designer to the fact that the load distribution factors of Section 1.3.1 (B) as
well as the minimum slab thickness and diaphragm provisions of Section
1.6.24 (C) and (F) of AASHTO Specification either cannot or should not
be applied to segmental bridges. Additionally, bridges which fall under the


Fig. 1.12 Typical cross section, lntracoastal Canal Bridge, Corpus Christi, Texas.

classification of segmental bridges, as used in this book, have traditionally

been erected using the cantilevering technique or other special erection
techniques which require special engineering analysis on the part of the
designer. The difference in design considerations between box-girder
bridges which are constructed in place on falsework and segmental bridges
which are erected with more sophisticated techniques will be brought to the
attention of the designer by this distinction in terminology.
‘Typical cross sections for segmental bridges are shown in Fig. 1.12
through 1.19. An examination of these cross sections will reveal that the
superstructures of the first three (Figs. 1.12, 1.13 and 1.14) are single-cell
tubes of constant depth. All three of these bridges were constructed using
precast segments. The superstructure of the Pine Valley Creek Bridge
(Fig. 1.15) has two single-cell constant depth tubular girders which are
structurally independent between the supports. The Pine Valley Creek
Bridge was constructed with the balanced cantilever technique with seg-
ments that were cast-in-place on traveling forms supported by the previ-
ously constructed superstructures. The Napa River Bridge superstructure
is shown in Fig. 1.16 from which it will be noted the depth ofthe two-celled
superstructure is variable. The Napa River Bridge, which was under
construction at the time this book was written, was being cast-in-place on
falsework using the balanced cantilever erection method. The cross sec-
tion of the B-3 Viaduct in Paris consists of two constant depth precast
2 0 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

39’-6” >1

Fig. 1 .13 Typical cross section, Bear River Bridge, Nova Scotia.

Fig. 1.14 Typical cross section, Bridge on U. S. Route 50 over the Vernon Fork df the
Muscatatuck River, Indiana.


Fig. 1 .15 Typical cross section, Pine Valley Creek Bridge, California.

single-cell tubular girders which are connected together transversely by a

common slab span. The typical section for the B-3 Viaduct is shown in Fig.
1.17. The variable depth superstructure of the Saint And& de Cubzac
Bridge is shown in Fig. 1.18 from which it should be noted that the upper
deck spans longitudinally between floor beams in the precast segments,
rather than transversely between the webs as is more commonly the case.
By far the most commonly used cross section in the construction of
segmental bridges is the single cell. The single cell sections may or may not
be connected together with a common middle slab as shown in Fig. 1.17.
The result is that the typical segmental bridge has considerably fewer webs
than has been the case with the typical box-girder bridge. It is expected the
number of webs used in box-girder bridges will decline in the future as a
result of bridge designers comparing the structural efficiency of the cross
sections of typical box-girder and segmental bridges. Cross sections with as
many as four cells and without deck overhangs have been constructed
using precast segments and the balanced cantilever erection method. The
Saint Cloud Bridge over the Seine in Paris, which is shown in Fig. 1.19, is
such a structure. The cross section of the Saint Cloud Bridge was not
selected for its structural efficiency but for its effect on the appearance of
the completed structure.

68’-0” c
9 N
7- %
4 1

Fig. 1 .16 Typical cross section, Napa River Bridge, California.

r 61’-6”




Fig. 1.17 Typical cross section, 63 Viaduct, Paris, France


Segmental superstructures also have high torsional stiffness, but not

generally as high as that of box-girder superstructures. The torsional
stiffness of the tubular girders is an important design consideration as will
be seen in Chapter 4.
Intermediate diaphragms are not used in superstrucures incorporating
tubular segmental girders. Provision of intermediate diaphragms in a
single-cell superstructure may or may not enhance the structural perfor-
mance of the structure (Ref. 27). Experience has shown that the torsional
stiffness of the tubular girders renders the provision of intermediate diap-
hragms unnecessary in superstructures consisting of two or more tubular
girders which are connected together by their top decks (Fig. 1.17). (See
Section 4.4).
Diaphragms, of solid or open construction, are almost invariably pro-
vided in segmental bridges at both the abutments and intermediate points of
support. These diaphragms are normally needed to transfer moments and
shears from the superstructure to the substructure. In small grade-separa-
tion structures, diaphragms have been omitted at the intermediate bents
and been provided at the abutments alone. Special construction details
must be provided for transferring loads and moments from the superstruc-

r2 27-3” 12’-9” _ _ 14’-6”


Fig. 1.18 Typical cross section, Saint Andre de Cubzac Bridge, France.

Fig. 1 .19 Typical cross section, Saint Cloud Bridge, Paris, France.

ture to the substructure when diaphragms are not provided at the inter-
mediate points of support.
Concrete slab bridges are normally used only for short spans because of
reasons of economy. Slab bridges can be constructed with precast pre-
stressed concrete components together with a cast-in-place topping or
might be entirely cast-in-place. Slab bridges require the least construction
depth of all the types of concrete bridges. Slab bridges are sometimes used
in lieu of other bridge types because of the minimum construction depth
associated with their use.
Typical cross sections of slab bridge superstructures are given in Figs.
1.20 and 1.21.
Slab bridges can be designed using the empirical relationships found in
Section 1.3.2 ofthe AASHTO Specifications, using the influence charts of
Pucher or Homberg (Ref. 17,18), or other methods of elastic analysis (Ref.

F 4

0 0 0 0 0 0 .o 0 0

Fig. 1.20 Typical cross section, slab bridge with precast soffit slabs. ’
I N T R O D U C T I O N 12 5

\ I
Fig. 1.21 Typical cross section, cast-in-place voided slab bridge.

28,29). Due to the relative flexibility of slabs, the transverse bending

moments which result from the application of concentrated loads must be
considered in their de.sign.
A bridge superstructure composed of longitudinal beams placed side-
bydside as shown in Fig. 1.22 is referred to as a multi-beam bridge in the
AASHTO Specification. Although bridges of this type could be formed of
I-shaped beams, they most commonly are made with precast, pretensioned
hollow-box beams as shown in Fig. 1.22.
The precast hollow-box beam possesses considerable torsional stiffness
and hence rotates only slightly if loaded eccentrically. It is essential that
the ends of the beams of multi-beam bridges be connected together with
transverse ties extending across the structure. It is also essential that
shear-transferring devices be provided between the individual beams. If
adequately connected together transversely by the provision of reinforcing
which extends completely across the structure together with shearconnec-
tors of sufficient size and shape, bridges of this type will behave structur-




Fig. 1.22 Typical cross section, multi-beam bridge.


ally in much the same manner as a monolithic box-girder bridge. If adequate

intermediate transverse ties are not provided, the joints between the beams
will tend to open under the application of concentrated live loads. This is
due to the lack of tensile strength along these joints together with the
flexural and torsional distortions of the beams. An excellent, although
relatively costly, means of attaining an adequate transverse tie consists of
providing post-tensioned tendons transversely at each end as well as at one
or more intermediate points along the span.
Empirical relationships for the distribution of live loads to multi-beam
bridges are given in Section 1.3.1 (D) of the 1974 AASHTO Interim
Specifications (Ref. 6). The theoretically correct distribution of live loads
to the beams of multi-beam bridges is dependent upon the adequacy and
efficiency of the shear connectors and transverse ties between the indi-

v 32’-0”


1 _ ,O

Fig. 1.23 Typical cross section, spread box beam bridge.

vidual beams. The empirical relationships found in the AASHTO Specifi-

cations are considered adequate for all commonly encountered conditions
of span, width and loading. A sophisticated elastic analysis of multi-beam
bridges would probably not result in sufftcient savings (if any) to justify the
extra engineering effort required to make the analysis.
Although the use of multi-beam bridges incorporating precast box beams
has been extensive in several parts of the country, they have not proved to
be economical in many other parts of the country.
The bridge superstructure formed of several precast box beams placed
parallel to each other at a spacing which exceeds their widths and which are
connected together with a cast-in-place deck, is referred to in the
AASHTO Specification as a spread box-beam bridge. A typical section for
a bridge of this type is shown in Fig. 1.23. Provisions for the distribution of
live loads to spread box-beam bridges are given in Section 1.6.24 (A) of the




Fig. 1.24 Typical cross section, bridge utilizing double-tee beams.

1974 AASHTO InterimSpecification. The empirical live load distribution

factors for spread box-beam bridges were derived by research conducted
on actual bridge superstructures.
End diaphragms must be provided with bridges of this type in order to
transfer lateral loads to the abutments and to prevent relative movement
between the beams at their ends. The provision of adequate intermediate
diaphragms in spread box-beam bridges reduces the live load induced
torsional rotations of the beams as well as the differential deflections
between beams and hence reduces the moments induced in the connecting
slabs. If intermediate diaphragms are not provided, moments will be in-
duced as a result of the torsional rotations and vertical deflections of the
beams when subjected to live load. These moments can be evaluated using
the methods described in Chapter 4 for segmental bridges having two

32’-0” A



Fig. 1.25 Typical cross section, bridge utilizing single-tee beams.


girders with a connecting slab. If one does not wish to evaluate these
moments, the empirical slab design relationships of AASHTO Section
1.3.2(C) can be used.
Bridges are sometimes constructed using precast concrete members of
types normally used in building construction. Typical sections for bridges
of these types are shown in Figs. 1.24, 1.25 and 1.26. These structures are
usually short-span structures and are provided with end diaphragms and
cast-in-place decks which, among other functions, transfer shear forces
between the girders. The structural behavior of these types of structures
is similar to either girder bridges without intermediate diaphragms or slab

0 0 Of0 0 of0 0 QfO 0 of0 0 of0 0 Of0 0 010 0 0

r- 3(r-0”

Fig. 1.26 Typical cross section, bridge utilizing hollow precast slabs.

1.7 Span Length vs. Bridge Type

The relative economy of the different types of prestressed concrete bridges

varies between the various parts of the country as well as with changes in
the general economy of the nation. Hence accurate specific conclusions
pertaining to the relative economy of bridge type versus span cannot be
made. One can, however, conclude from a review of practical applications
that bridges of short span will probably be most economical using slab-type
structures. Moderate spans are most economically done as girder-type
bridges while the longer span structures are generally box-girder or
segmental-girder bridges. The longest concrete bridge spans have utilized
cast-in-place segmental-type superstructures.
2 Girder

2.1 Introduction

Girder bridges of two commonly used types were illustrated in Chapter 1.

Typical cross sections for girder bridges of these types are shown in Figs.
1.7 and 1.8. It was explained that the negligible torsional stiffness of girders
having I-shaped or T-shaped cross sections necessitates the provision of
intermediate diaphragms in all girder bridges except for those of relatively
short span. It will be shown in this Chapter that the torsional flexibility of
the beams in girder bridges has great bearing on their response to load.
The designer of a girder bridge must determine the portion of the dead
and live load that is distrubuted to each of the individual girders in the
bridge. This can be done by elastic analysis or by the use of empirical
methods. The most exact methods of elastic analysis involve the use of
large electronic computers and sophisticated programs. The cost of sophis-
ticated analyses of this type can not be justified except for structures of
major importance. For the more commonly encountered conditions, ap-
proximate methods ofelastic analysis yield adequate results. The empirical
methods have proven to yield results which are generally satisfactory from
a performance standpoint and are economical for structures of small or ,
modest size.
30 ) H I G H W A Y B R I D G E S U P E R S T R U C T U R E S


Fig. 2.1 Typical cross section of a multi-girder bridge superstructure.

Consider a multi-girder bridge superstructure, such as is shown in Fig.

2.1, which is intended for use on a span of 120 feet. If the girders are
assumed to be without torsional stiffness and if intermediate diaphragms
are not provided to connect the girders transversely, the lateraldistribution
of loads applied to the deck is accomplished by the deck slab. For the
application of the dead load due to the safety walks and bridge railing, the
structure would be expected to deflect as shown in Fig. 2.2 from which it
will be seen that the exterior girders are required to carry more of the load
than are the interior girders. For the application of concentrated loads on
the deck, the structure would deform as shown in Fig. 2.3 from which it will
be seen that the interior girder nearest to the point of application of the
load, carries the largest portion ofthe load. From this it should be apparent
that transverse flexibility is an important design consideration for girder
bridges that are without intermediate diaphragms.
If an intermediate diaphragm is provided at midspan of the bridge of Fig.
2.1, it may be as shown in Fig. 2.4. It can be shown that the stiffness of the


J 1

Fig. 2.2 Typical cross section of a multi-girder bridge which does not have intermediate
diaphragms and which is subject to the loads of safety walks and railings.

Fig. 2.3 Typical cross section of a multi-girder bridge which does not have intermediate
diaphragms and which is subject to a single eccentrically-applied concentrated

diaphragm is great in comparison to that of the girders for these conditions

(span 120 feet, distance between exterior girders, 28 feet, and depth of
diaphragm of the order of that of the girders). For this example, the
moment of inertia of the intermediate diaphragm is of the order of 533,000
in4 while that of the deck slab alone is 244 ina per foot. If one were to
assume the entire 120 feet of bridge deck effective in distributing loads
transversely (it obviously is not) the stiffness of the deck slab would be of
the order of one-eighteenth as much as that of the intermediate diaphragm.
Hence the application of the safety walk and railing dead loads result in a



23’-(y’! @ 7’~Q” = 14’4”

Fig. 2.4 Typical cross section of a multi-girder bridge superstructure which has an inter-’
mediate diaphragm.
3 2 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

Fig. 2.5 Deflection of a multi-girder bridge having an intermediate diaphragm due to the loads
of safety walk and railings.

vertical deflection with virtually no transverse bending deflection as shown

in Fig. 2.5. The application of an eccentrically applied concentrated load
over the midspan diaphragm, will result in a deformation as shown in Fig.
2.6 from which it will be seen the transverse flexural deformation is virtu-
ally nonexistent and the exterior beam’on the loaded side carries the

1 #
J 1

Fig. 2.6 Deflection of a multi-girder bridge having an intermediate diaphragm due to a sihgle
eccentrically-applied concentrated load.

greatest portion of the load. The provision of a well designed intermediate

diaphragm obviously has important effects on the distribution of loads in a
girder bridge, such as this one.
For girder bridges which are wide in comparison to their span, even
though they may have intermediate diaphragms, transverse elastic deflec-
tion resulting from the application of concentrated loads can be significant
and should not be ignored. Such superstructures act in a manner that is not
unlike the action of an elastic plate that is supported on two edges (their
Consider the special case of a long span bridge which incorporates only
two torsionally flexible girders as shown in Fig. 2.7. Concentrated loads
applied to the deck of such a structure are distributed to the girders as if the
deck were a simple span over two supports. The provision of an inter-
mediate diaphragm at midspan of the structure will not change the distribu-

Fig. 2.7 Cross section of a bridge containing two longitudinal girders.

l tion of loads to the girders. The provision of additional intermediate dia-
phragms will not change the distribution of loads to the girders nor change
, the shape of the deflected girders either transversely or longitudinally.
I From this simple example it is seen that the bridge superstructure of this
type is a special case in which the distribution of loads to the girders can be
determined with the basic rules of statics.
/. In employing the sophisticated methods of analysis, it is suggested that
the procedure be to develop influence lines for the various critical mo-
ments, shears and deflections required to insure a safe and serviceable
I structure rather than attempting to analyze the structure as a whole for the
i many combinations of live loading that could be imposed upon it. With this
3 4 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

approach, influence lines for moments or shears at various locations in the

structure, similar to those which are shown in Figs. 4.42 and 4.43 for the
moments in the connecting slab of the two-tubular girder superstructure,
can be developed with minimal computer cost.
An approximate method for the elastic analysis of girder bridges has
been proposed by Courbon (Ref. 31). This method is applicable only to
girder bridges which have intermediate diaphragms of adequate design (see
Section 2.3). When the span length of the girders is equal to or greater than
twice the distance between the fascia girders and the depth of the dia-
phragm is of the same order as the girder depth, the diaphragm may be
considered to be infinitely stiff in comparison to the girders. It is claimed
that under these conditions this assumption will result in an error of less
than four percent in the live loads computed for any ofthe principal beams.
The method proposed by Co&bon can be understood by considering a
typical girder bridge superstructure composed of several parallel beams,
each of which has a constant moment of inertia, and one intermediate
diaphragm of infinite stiffness at midspan. The torsional stiffness of the
beams is considered to be negligible. If a load is placed eccentrically over
the diaphragm as shown in Fig. 2.8, the system will deflect downward with
the.exterior girder on the same side as the eccentric load (1) having the
largest deflection whereas the other exterior girder (5) will have the
smallest deflection. The interior girders will have deflections which vary


Fig. 2.8 Deflection at midspan of a multi-girder bridge which has an adequate intermediate
diaphragm and which is subject to a single eccentrically applied concentrated load.

linearly between those of the exterior girders. It should also be apparent

that the elastic curve of each of these members would be the same even if
intermediate diaphragms existed at other locations in the span in addition
to the one at midspan. If these other diaphragms did exist, under the
condition of loading described above, they would not be transmitting loads.
The reaction of the diaphragm (Ri) on each girder can be evaluated by
first determining the elastic center of the system. The elastic center is
defined as the point measured along the diaphragm where the summation of
the products of the moments of inertia of the girders and their distances
from the elastic center are equal to zero. If the moment of inertia of beam i
is represented by Ii and it is at a distance of xi from the origin, the elastic
center is computed from:

If a load P is placed upon the diaphragm at eccentricity e, the reaction of the

diaphragm on beam i will be:



,i=xL[1+23$L] (2.3)

In the case of a system composed of n beams equally spaced at intervals

of S feet each of which has the same moment of inertia, the reaction on
beam i is:
Pi=$[ 1 +6( ‘;z\-~~~ ) :I (2.4)

In Eq. 2.4, the beam number i is to be taken with beam number 1 being on
the same, side of the axis of symmetry as is the eccentricity. This is
illustrated in Fig. 2.8
A comparison of the results obtained with the approximate method just
explained to that found with a finite element analysis for the superstructure
of Fig. 2.4 is shown in Fig. 2.9. It is apparent that the approximate analysis
yields results of sufficient accuracy for normal bridge design work. It is
important to note that each method of analysis reveals that the exterior
girder is subjected to a considerably greater portion of the load than the \
interior girders.
3 6 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

100 KIPS

0 - I’
3.125 -

:O 250.
$o.375---- a---. - - ----_-- - -


5 4 3 2 1


100 KIPS


f 0.75 .
$ , M)(1 APPROXI dATE A N A L Y S I S - ’

5 4 3 2 1

Fig. 2.9 Deflection of the bridge having the cross section of Fig. 2.4 as determined by the
approximate and finite element methods for (a) a single concentrated load at tQe
center, and (b) a single concentrated load with an eccentricity of seven feet.

The distribution of dead and live loads to interior and exterior beams is
covered in Section 1.3.1 of the AASHTO Specification. The wheel load
distribution factors for interior beams that have been included in the
AASHTO Specification since 1949 are shown in Table 2.1. Note that the
distribution factors vary from S/5.0 to S/6.0 for torsionally flexible beams
of different types. The reason for this variation is not clear and cannot be
rational. The wheel load distribution factors for interior beams supporting
concrete decks, as contained in the 1973 edition of the AASHTO Specifi-
cation, are given in Table 2.2.
Using the empirical wheel load distribution factors from the AASHTO
Specification, the portion of a wheel load assumed to be imposed upon an
interior concrete T-Beam and a prestressed concrete girder are as shown in


I - -


// / ’
1/ I 2 4 6
6 10 12 14

Fig. 2.10 Comparison of the empirical wheel load distribution factors for an interior girder as ,
provided in the 1973 AASHTO Specification for a concrete T-beam (S/6.0) and a
prestressed concrete girder (S/5.5).

Table 2.l-Live Load Sanding Moment Distribution Factors From AASHTO Specifioation For
Concrete Floors Supported by Various Types of Beams

r Type of Beam 1
Yew steel zalcrets hcretc j
spzed I-Beams shingem
s xrlnger
1949 Y5.0
1953 Y5.0
1957 Y5.5 Y5.0 Y5.0
I%I Y5.5 Y5.0 Y5.0 Y7.0
I%5 Y5.5 S/6.0 Y5.5 Y5.0 Y7.0
1969 Y5.5 Y6.0 Y5.5 Y5.0 Y7.0
1973 Y5.5 Y6.0 Y5.5 Y5.0 Y7.0 .
l *
1974 Y5.5 Y6.0 s15.5 Y5.0 s/7.0
‘Spread box beam superstructures are covered m SectIon 1.6.24(A) of the AASHTO Specification by speaal empirical
“Multi-beam bridges, formed of precast box girders placed side-by-side. were treated as slabs prvor lo the 1974 Interim
Speafication. They are currently covered in Sedion 1.3.1 (D) of the 1974 Interim Speclcation.


Bridge designed Bridge designed

Kind of Floor for one for two or more
traffic lane traffic lanes

On Steel I-Beam Stringers
and Prestressed Concrete
Girders . . . . . . . . . . . . . . . . . . s/7.0 SIC.5
IIf S Exceeds IO’ us6 If S exceeds 14’ use
footnote footnote
On Concrete T-Beams __.__.. S/6.5 S/6.0
If S exceeds 6’ use If S exceeds IO use
footnote footnote
On Timber Stringers . . ..____.. S/h.0 s/s.0
If S exceeds 6’ use If S exceeds IO’ use
footnote footnote
Concrete box girders S/8.0 s/7.0
If S exceeds I?’ use If S exceeds 16’ use
footnote footnote
S = average stringer spacing in feet.

Footnote: In this case the load on each stringer shall be the reaction of the wheel
loads, assuming the flooring between the stringers to act as a simple

Fig. 2.10. Note that even though the stiffnesses of the two beams may be
identical, the prestressed girder is assumed to carry more load than the
The 1969 edition of the AASHTO Specification provided that outside or
exterior beams of girder bridges be designed for the dead load of the deck
which they support. This is normally taken to mean the portion of the deck
load that would be imposed upon the exterior beam if the deck slab were
simply supported at the frost interior beam and the exterior beam. In
addition, the 1969 AASHTO Specification provided that the loads from
curbs, railings and wearing surfaces could be considered to be equally
distributed to all roadway beams if these loads were imposed after the deck
slab had been cured. The live load to be carried by the exterior beam was
specified as the reaction ofwheel load or loads determined by assuming the
deck slab to act as a simple span between the exterior and fust interior
Using the provisions of the 1969 edition of the AASHTO Specification,
exterior beams were frequently designed for less live load than were the
interior beams. A rational elastic analysis of girder bridges, as has been
described above, reveals that the opposite situation might actually exist.
The exterior beams are subject to greater stresses due to live load than are
the interior beams if the intermediate diaphragms are of adequate design.
A more recent edition of the AASHTO Specifications states that the
exterior beams of girder bridges shall in no case have less carrying capacity
than an interior beam. The term “carrying capacity” is not defined and
could be subject to different interpretations.
The number of traffic lanes for which a bridge is to be designed in
accordance with the provisions of the AASHTO Specifications is covered
by Article 1.2.6 as revised by the 1974 Interim Bridge Specifications. The
provisions of this Article are as follows:
The lane loading or standard truck shall be assumed to oc-
cupy a width of ten feet.
These loads shall be placed in 12-foot wide design traffic
lanes, spaced across the entire bridge roadway width, in num-
bers and positions required to produce the maximum stress in
the member under consideration. Roadway width shall be the
distance between curbs. Fractional parts of design lanes shall
not be used. Roadway widths from 20 to 24 feet shall have two
design lanes each equal to one-half the roadway width.
The lane loadings or standard trucks having a IO-foot width
shall be assumed to occupy any position within their individual
design traflic lane, which will produce the maximum stress.
It is interesting to compare the results one obtains for the portion of a ’

wheel load carried by a bridge girder using the AASHTO empirical coeffr-
cients of Table 2.1, to those obtained from the approximate elastic analysis
of Courbon. Using the notation shown in Fig. 2.11 and the provisions of
AASHTO Article 1.2.6, it can be shown that the maximum eccentricity of
the live load on a bridge that is 24 feet wide or wider is as follows:
emax = W- -6N+l (2.5)
in which N is the number of design traffic lanes on the bridge and W is the
width between curbs or railings. Using the Eqs. 2.4 and 2.5 for various
combinations of bridge roadway width W, number of traftic lanes N, and
number of beams n, one obtains the results shown in Table 2.3. An
examination of these results will show that the AASHTO coeffkients are
more conservative for bridges designed for two design traffic lanes than for
bridges of greater width. The results summarized in Table 2.3 also illustrate
that the AASHTO empirical coeffkients do not reflect the effects of
number of design traffic lanes on a superstructure nor the effect of load
intensity reduction for multi-lane structures.
Some engineers feel that the design traffk lanes should be ten feet wide
rather than twelve feet wide as provided by the AASHTO Specifications.

W=28’0” P



Fig. 2.11 Cross section of a bridge used to illustrate positioning of live loads as provided /)y the
AASHTO Specification.
G I R D E R B R I D G E S 14 1

12’ Traffic Lanes

Portion of a Wheel Load

Interior Stringer AASHTO
n .Exterior Prestressed
W N number S Stringer Concrete Concrete
Ft. each Pt. of beams Ft. Eq. 2.4 T-Beam Girder

20 1 5 4 6.33 0.97 1.06 1.15

2 0 4 6.33 1.00 1.06 1.15

24 I 7 4 7.67 1.05 1.28 1.39

2 1 4 7.67 1.16 1.28 1.39

36 1 13 6 7.00 0.86 1.17 1.27

2 7 6 7.00 1.24 1.17 1.27
3 1 6 7.00 1.12s 1.17 1.27

48 1 19 7 7.83 0.81 1.31 1.42

2 13 7 7.83 1.28 1.31 1.42
3 7 7 7.83 1.43* 1.31 1.42
4 1 7 7.83 1.25* 1.31 1.42

10’ Traffic Lanes

Portion of a Wheel Load

Interfor Stringer AASHTO
I I I I n I Exterior I Prestressed
W N number S Stringer Concrete Concrete
Ft. each i. of beams Ft. Eq. 2.4 T-Beam Girder

20 1 5 4 6.33 0.97 1.06 1.15

2 0 4 6.33 1.00 1.06 1.15

24 1 7 4 7.67 1.05 1.28 1.39

2 2 4 7.67 1.31 1.28 1.39

36 1 13 6 7.00 0.86 1.17 1.27

2 8 6 7.00 1.32 1.17 1.27
3 3 6 7.00 1.37* 1.17 1.27

48 1 19 7 7.83 0.81 1.31 1.42

2 14 7 7.83 1.34 1.31 1.42
3 9 7 7.83 1.60* 1.31 1.42
4 4 7 7.83 1.58* 1.31 1.42

* These factors could be reduced to 90% and 75% of the values obtained from Eq. 2.4 for three
and four design traffic lanes, respectively, in compliance with Section 1.2.9 of the
AASHTO Specifications.

Using the design traffic lane width of ten feet rather than twelve feet, the
relationship for maximum live load eccentricity becomes:

With this value of eccentricity, one obtains wheel load distributions to the
girders as shown in Table 2.4.
From the preceding discussion one must conclude that the provisions of
AASHTO regarding the distribution of live load to interior girders of
concrete bridges are conservative for bridges having two design traffic
lanes and are progressively less conservative as the bridge roadway width
increases and the width of the design traffic lane decreases. They do not
vary in the same manner as one obtains from an approximate elastic
analysis. It can also be shown they do not agree with the results one obtains
from a rigorous elastic analysis.

2.2 Girder Design

Once the distribution of the loads to the girders, as was discussed in

Section 2.1, has been determined, the design of the girders can be under-
taken. Two procedures are equally acceptable. In the fust, the girders are
proportioned for design loads and checked for their adequacy under the
appropriate serviceability requirements. The second procedure consists of
designing for service loads by elastic analysis after which the strength of
the member is checked to confirm compliance with the minimum strength
requirements of the design criteria being used. The most complete single
source of data relative to the service and strength design requirements for
prestressed concrete bridges will be found in reference 32.
Girder bridges require a relatively great structural depth to achieve a
specific span. The depth-to-span ratios normally employed in girder
bridges ranges from 1 in 16 to 1 in 20. The data presented in Appendix B
relative to the dimensions of the more commonly used precast prestressed
concrete bridge girders includes the span ranges on which the various
sections are commonly used. The girders are normally used at spacings of
from 6 to 9 feet with the closer spacings being used with the lower depth-
to-span ratios. The fact that girder bridges do require relatively great depth
has resulted in their use not being economical or perhaps even feasible in
certain applications where vertical clearance requirements have been criti-
Proportioning prestressed concrete bridge beams for optimum structural

efficiency involves two basic considerations which may not be obvious.

The first of these is the amount of the total dead load moment that is acting
upon the section under consideration at the time of prestressing. The
second is the ratio of the dead load moment to the total moment for which
the section must be designed. These factors are best illustrated by consid-
ering several examples. Before so doing, two fundamental axioms of struc-
tural engineering should be mentioned. These are: (1) Flexural members
of relatively short span are generally more critical with regard to shear
stresses whereas members of relatively long span are generally more
critical with regard to flexural stresses, and (2) For flexural members of
relatively short span, live load generally constitutes a larger portion of the
total load for which the member must be designed than does the dead load,
while the opposite is true for flexural members of relatively long span.
With the foregoing in mind, it is interesting to consider how the optimum
shape of a simply supported prestressed concrete bridge girder will vary
with span length. The girder for a short span requires a relatively thick web

L zq

II Z-4” 4
Fig. 2.12 Cross sections of the AASHTO-PCI Standard Bridge Beams.
44 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

because of the greater importance of shear stresses. In addition, because

the dead load moment is relatively unimportant in comparison to the total
moment, a large bottom flange is required to “store” the prestressing force
until needed to resist the transient live loads. The size of the top flange of a
short-span bridge beam is rarely if ever critical because the bridge deck,
whether cast monolithically with the beam or separately, is almost invari-
ably adequate to sustain the moments at service and design levels.
For simple prestressed concrete bridge beams of long span, flexure is the
primary design consideration. Because the bulk of the moment is due to
dead loads, most of which may be acting at the time of prestressing, a large
bottom flange is not required. A substantial top flange is needed for both
service and strengthconsiderations. Shear stresses are relatively unimpor-
tant in the design of long span prestressed concrete bridge beams.
Further consideration of the above will reveal that precast prestressed
girders of optimum design will have different shapes for different spans.
This is illustrated in Fig. 2.12. Long, simple-span cast-in-place members of
prestressed concrete which have T-beam cross sections similar to that
shown in Fig. 1.7, can be shown to be efficient.
In the design of superstructures incorporating precast members which
are designed to act compositely with a cast-in-place deck, the elastic
analysis for the composite section should be based upon section properties
computed with due regard to the difference in the elastic modulii of the two
concretes. In the design of large post tensioned concrete beams of T- or
modified I-shaped cross sections, the space occupied by the ducts for the
post-tensioning may have significant influence on the initial concrete
stresses in the member. Hence, it is recommended that large post-
tensioned girders be designed taking into account the effects of the net,
transformed and composite section properties (Ref. 33).
Dimensions for typical precast prestressedconcrete bridge beams which
have found wide use in the United States are given in Appendix B.

2.3 Intermediate Diaphragms

An intermediate diaphragm has an important funtion to perform in insuring

the structural integrity of a prestressed concrete girder bridge which has
more than two longitudinal girders. There are no specific requirements in
the AASHTO Specifications relative to the design of intermediate dia-
phragms for concrete bridges. This is unfortunate because many details
which are commonly used for intermediate diaphragms with precast pre-
stressed concrete girders are poor and not adequate at loads approaching
the flexural capacity of the girders (Ref. 34, 35,36).

As previously stated, an intermediate diaphragm should be provided at

midspan of each span of a girder bridge. Little if any benefit is to be gained
by placing additional intermediate diaphragms between the supports and
midspan because they have little if any additional effect in improving the
deflected shape of the girders. The intermediate diaphragms should be
designed for the reactions computed with Eq. 2.4 unless more exact
methods of analysis are used. Care must be given in detailing the connec-
tion of the diaphragm to the exterior girders. The continuity of the dia-
phragm flexural reinforcement between the exterior girders must be as-
sured if the diaphragm is to function properly. Adequate shear reinforcing
must be provided. One of the best methods of constructing intermediate
diaphragms not frequently used, although it has been used in the past, is to
post-tension the diaphragm as shown in Fig. 2.13. Unfortunately, post-
tensioned diaphragms are more costly than those of reinforced concrete
and it is this that undoubtedly accounts for their no longer being used.



Fig. 2.13 Details of a post-tensioned intermediate diaphragm.


Fig. 2.14 Intermediate diaphragm composed of structural steel shapes.
4 6 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

A novel approach to providing stiff intermediate diaphragms in girder

bridges that has been successfully used on bridges constructed for the
lumber industry in the State of Washington is shown in Fig. 2.14 (Ref. 37).
These diaphragms consist of structural steel sections installed by connect-
ing them to small pieces of steel cast into the precast beams. Diaphragms of
this type can be installed much more rapidly than the more conventional
cast-in-place concrete diaphragms. They do, however, require mainte-
nance as do bridges incorporating steel beams and trusses.

2.4 Decks for Girder Bridges

In the design of the deckfor a girder bridge, a distinction must be made

between bridges which have flexurally-stiff intermediate diaphragms and
those which do not. As previously pointed out, the individual bridge
girders of normal configuration (i.e. I-shaped or T-shaped beams) have
virtually no torsional stiffness and hence do not afford rotational restraint
to the decks which they support. Because the decks themselves are rela-
tively flexible in comparison to the girders, differential girder live load
deflection must be considered for bridges which do not have adequate
intermediate diaphragms.
In view of this, it is recommended that the decks of girder bridges
containing more than two girders be designed using the empirical relation-
ships of the AASHTO Specification in cases where the girders are torsion-
ally flexible and well designed flexurally stiff intermediate diaphragms are
not provided between them.
In the case of multi-girder bridges incorporating girders which are tor-
sionally flexible but connected by well designed flexurally-stiff inter-
mediate diaphragms, it is recommended the decks be designed as being
continuous over the supporting girders without differential settlement bet-
ween supports. The analysis should be made on an elastic basis using the
Pucher (Ref. 17) or Homberg (Ref. 18) charts or other sophisticated
methods of analysis. The girders should not be considered as providing
rotational restraint to the deck.
As a means of illustrating the effects of differential girder deflection on
the flexural stresses in the deck of a bridge incorporating precast I-girders,
the results of a finite element analysis for the cross section of Fig. 2.1 are
summarized in Fig. 2.15. The analysis was made for the bridge having a
simple spanof 120 feet and having one intermediate diaphragm at midspan.
Two concentrated loads of 100 kips were applied to the bridge. The loads
were applied over the middle girder 30 feet from each end of the bridge as

100 KIPS 100 KIPS




100 KIPS

4 3 2 1

4 ),G,, = -7.47 x 10-b RADIANS

(c) G,, = 0 RADIANS f

Fig. 2.15 Study of secondary stresses in the deck of a girder bridge; (a) loading arrangement, ,
(b) def!sction at point of application of a 100 kip load, and (c) elastic curve of the deck
between points 3 and 4.
48 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

shown in Fig. 2.15(a). Under this condition of loading, the maximum

flexural stress in the slab occurs at the point of application of the load and is
of the order of + 320 psi. When it is realized that the 100 kip load is of the
order of five times the design wheel load (including impact) used with HS
20-44 loading it becomes apparent the live load stresses due to differential
girder deflection in structures of this type are of secondary magnitude and
can safely be ignored in practice providing the design criteria being
employed reasonably reflects the actual wheel loads to be imposed on the
As pointed out above, bridge superstructures which contain onIy two
torsionally flexible longitudinal girders constitute a special case. The decks
for bridges of this type (see Fig. 2.7) should be designed as simply sup
ported spans because the torsionally flexible girders provide virtually no
joint restraint at all. The designer has the option of designing decks of this
type with the influence charts of Pucher (Ref. 17) or Homberg (Ref. 18) or,
of course, with the empirical coefficients of the AASHTO Specification.
The use of the influence charts will yield more rational results.
When torsionally stiff girders are used, it is recommended that special
methods of analysis similar to those discussed in Section 4.4 be employed.

2.5 Continuity
Prestressed concrete girder bridges utilizing cast-in-place T-beams are not
a commonly used form of construction. Cast-in-place reinforced concrete
T-beam bridges are rather frequently used and it is not uncommon for them
to be made continuous over three or more supports. Concrete T-beams
have found greatest use in bridges of medium span length and this probably
accounts for the rather infrequent use of this shape with prestressing. The
design of a T-beam bridge using prestressed reinforcement is straightfor-
ward and only the computation ofthe secondary moment resulting from the
prestressing is a factor that is markedly different from the design proce-
dures that must be followed with the two different materials.
Continuity has often been established in bridges incorporating precast
prestressed concrete beams. This is normally accomplished through the
provision of reinforcing steel in the cast-in-place deck which is placed over
the precast girders. In this mode of construction the major portion of the
dead load is carried by the precast beams acting as simple beams. The dead
load of railings, sidewalks, wearing surfaces and other superimposed loads
are carried by the continuous structure, as are the live loads. In the design
of bridges of.this type, positive moments may occur at the interior supp,orts
due to the effects of temperature as well as creep and shrinkage of the

concrete. Special consideration should be given to these effects. In addi-

tion, due to the special nature of the stresses in the prestressed concrete
girders and cast-in-place diaphragms for bridges of this type, certain devia-
tions from the usual allowable concrete stresses at service and design loads
seem to be appropriate. These special requirements appear in reference 38.

2.6 Overhanging Beams

Occasionally the bridge designer is faced with providing a high, long,
channel span over a waterway in order to provide horizontal and vertical
clearance for navigation. A traditional method of accomplishing this in
steel bridges has been with the use of a span suspended from two overhang-
ing beams. This mode of framing, which is illustrated in Fig. 2.16, presents
certain difficulties when done with i-shaped prestressed concrete girders.
Most of these difficulties are in the form of construction sequences and
methods that must be followed. The effort required to design a bridge of
this type is aggravated by the need to establish the construction sequence
that must be followed. These difficulties can generally be eliminated or
minimized by the use of a cast-in-place box girder section for the overhang-
ing beams, in lieu of the use of I-shaped girders.


66’ I- 145’ -L 6 6 130
I- I- I-

Fig. 2.16 Elevation of a bridge with two overhanging beam spans and a suspended span.

Problems which have been encountered in the use of I-shaped pre-

stressed concrete girders in bridges with overhanging beams include the
1. Longitudinal prestressing of the girders frequently has to be done in
several stages, as the construction progresses, in order to control
initial concrete stresses in the structure.
2. Temporary prestressing tendons frequently must be used to control
stresses in the girders at certain stages in the construction.
3. The sequence of placing concrete in the end diaphragms, inter-

mediate diaphragms and deck slabs must take into account the effects
of elastic and non-elastic shortening of the concrete.
4 . Elastic stability of the girders (buckling) must be given consideration.
5. Erection procedures that can be used with the suspended spans may
be restricted due to the stresses that could be imposed on the over-
hanging beams.
6 . Excessive principal tensile stresses sometimes occur in the overhang-
ing portions of the beams prior to erection of the suspended spans.

Each of the above should be given careful consideration before this mode
of framing is adopted for use.

2.7 Construction Details

Multi-span girder bridges which incorporate precast prestressed concrete

members are frequently designed with all spans being of equal length. This
is done for the purpose of having all beams as identical in detail as possible.
The cost of the precast members will generally be lower when the members



Fig. 2.17 Recommended detail for deck-to-girder connection.

I 51

are of the same length and very similar in detail. With equal span lengths,
the negative moments at the first interior supports will be greater than those
at the other interior supports. This is not generally a serious problem
because the negative moments result only from superimposed dead loads
and live loads and, particularly for the longer span structures, are of less
importance than the primary dead load which is carried by the precast
members acting as simple beams.

Fig. 2.18 Detail for deck-to-girder connection which is not recommended.

Multi-span girder bridges of the cast-in-place T-beam type would nor-

mally, if possible, be proportioned with the end spans being about 80% as
long as the interior spans. Where possible, this arrangement results in a
well balanced distribution of moments.
Precast prestressed concrete beams frequently exhibit significant and
variable camber. If not anticipated in the detailing of the bridge, this factor
can cause considerable difftculty in the construction of the bridge. The
commonly used detail shown in Fig. 2.17 is recommended as a means of
accommodating camber and variations in camber. The additional slab
depth specified at the centerline of the bearings (1 SO inches for Fig. 2.17)

/ /
----__--__------ ----__--_______

------_____-- -------------
/ / /
Fig. 2.19 Preferred layout of diaphragms.

can sometimes be made less in structures which are on “humped” vertical

curves. For structures which are on “sagged” vertical curves, greater
thickening of the deck may be required over the girders at the centerline of
the bearings. The detail in Fig. 2.18 is occasionally seen and is not recom-
Diaphragms between the girders, including those at the ends as well as
those which are at midspan, should be placed along the skew on skewed
bridges, as shown in Fig. 2.19, rather than perpendicular to the girders in a
stepped layout as shown in Fig. 2.20. With the diaphragm at midspan along
the skew, points of equal flexural stiffness in adjacent girders are con-
nected to each other and continuity of the diaphragm flexural reinforce-
ment can be achieved. With stepped intermediate diaphragms points of
unequal stiffness are connected, continuity of the diaphragm flexural rein-

Fig. 2.20 Diaphragm layout which is not recommended.

Fig. 2.21 Poway Road Overcrossing of State Route 163, San Diego, California. (Courtesy of
California Department of Transportation.)

Fig. 2.22 Harbor Drive Overcrossing, San Diego, California. (Owner: City of San Diego.)
’ \
\ \
’ \’

Fig. 2.23 Los Penasquitos Creek Bridge, San Diego, California. (Courtesy of California
Department of Transportation.)
Fig. 2.24 West Mission Bay Drive Bridge, San Diego, California. (Owner: City of San Diego.)

Fig. 2.25 Pacora River Bridge, Republic of Panama. (Owner: Republic of Panama.)

forcing is not possible and torsional stresses are induced in the girders near
the diaphragms. It must be recognized that the diaphragm connection
details for bridges incorporating precast girders can become complicated in
detail and difftcult to construct along the skew as shown in Fig. 2.19,
especially for a structure having a large skew.
Girder bridges incorporating”precast prestressed concrete girders have
been constructed with bent caps located below the superstructure as well
as within the depth of the superstructure. The former type of cap construc-
tion is shown in Figs. 2.21 and 2.22. Although attractive appearance can be
achieved with this mode of framing, many engineers prefer the appearance
that is achieved when the pier or bent cap does not protrude below the
superstructure as shown in Figs. 2.23 and 2.24.
A cast-in-place post-tensioned T-beam bridge having variable depth is
shown in Fig. 2.25. This structure, has a main span of 200 feet. A sus-
pended channel span of 120 feet is supported on overhanging beams can-
tilevering 40 feet from each pier.
The details ofconstruction for prestressed concrete girder bridges varies
throughout the country. Details commonly used in each locality should be
determined from local contractors and producers of precast concrete. If
the details which are commonly used are considered adequate, their use
may be less costly than new and unfamiliar details. Other sources of
standard details include the Prestressed Concrete Institute*, the various
State Highway Agencies and the Federal Highway Administration.**

*Prestressed Concrete Institute, 20 N. Wacker Drive, Chicago, Illinois 60606.

**Federal Highway Administration, 400 Seventh Street, S.W., Washington, D.C. 20390
. 3 Box-Girder

3.1 Introduction

Typical cross sections for concrete box-girder bridges of the types which
are commonly used in the United Sates are shown in Figs. 3.1, 3.2, and
3.3. Bridges of these types, in both reinforced and prestressed concrete,
have been widely used in the western part of the country, although their
use has not been limited to this region.
The relative economy of the box-girder bridge has contributed greatly
to its popularity as has its relatively slender and unencumbered appear-
ance. Some proponents of the box-girder bridge have claimed its smooth
sofftt to be very desirable in urban areas for reasons of esthetics. The
structural simplicity of the box-girder bridge, particularly in continuous
structures of medium to long spans, has been well demonstrated. The
efftciency of the cross section for positive and negative longitudinal
bending moments as well as for torsional moments is apparent even to
the casual observer. Especially important advantages for this mode of
concrete bridge construction include the low depth-to-span ratio that can
be economically achieved together with the ease with which variable,
conditions of bridge width, superelevation and curvature, both vertical
58 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S


TX ; 1: I 1 d


Fig. 3.1 Cross section of a box girder bridge with vertical exterior webs.

Fig. 3.2 Cross section of a box girder bridge with inclined exterior webs.

Fig. 3.3 Cross section of a box girder bridge with curved exterior webs.

and horizontal, can be accommodated. Variable superstructure depth

can also be accommodated with relatively little difficulty.

3.2 Flexural Analysis

The torsional stiffness of the typical “closed” box-girder bridge cross

section as contrasted with that of the “open” I- or T-shaped girder sec-
tion is very significant. Also important is the transverse flexural stiffness
of the box-girder cross section. Each of these should properly be taken
into account in the flexural analysis of bridges of this type. The torsional
stiffness virtually eliminates significant transverse rotation of the section
due to eccentrically applied loads of the magnitudes used in contempor-
ary highway bridge design. This characteristic significantly differentiates
a box-girder bridge from a girder bridge of similar physical proportions.
The relatively large transverse flexural stiffness, in combination with the
torsional stiffness, eliminates the need for intermediate diaphragms
which are so important in girder bridges.
To illustrate the action of a box-girder bridge under concentric and
eccentric loading, the bridge shown in Fig. 3.4 was analyzed using the
finite element method as well as the familar flexural theory. In the latter
case, the torsional stiffness of the gross cross section was assumed to be
completely effective in resisting rotation due to eccentrically applied
loads and the transverse cross section was assumed to be infinitely stiff
flexurally. The analysis was made for the condition of no intermediate
diaphragms and a simple longitudinal span of 120 feet. The results of the
deflection analysis for the conditions of a single concentrated load of 100
kips applied concentrically and eccentrically are illustrated in Figs. 3.5
and 3.6, respectively. An examination of these figures will reveal the
approximate flexural theory yields results that are quite close to that

Fig. 3.4 Cross section of a simple span box girder bridge used in the comparison of the \
results obtained with the approximate and finite element methods of analyses.

100 KIPS

5 0.125

p 0.250
x - - - -
- - - - - -
f - -- - - /---


0.500 ’ 3
5 4 3 2 1

Fig. 3.5 Transverse deflection of the bridge cross section shown in Fig. 3.4 due to a single
concentrically-applied load of 100 kips as predicted by the approximate and finite
element methods.

obtained by the more sophisticated (and costly) finite element method of

analysis. The approximate method obviously yields results that are suf-
ficiently accurate for normal design work.
When one considers the fact that the load of 100 kips represents more
than a single HS 20-44 truck applied at a single point, rather than being
distributed in six concentrated loads of varying magnitudes spread over
a width of six feet and a length of 28 feet, it becomes apparent the ap-
proximate analysis yields conservative results for the analysis of lon-
gitudinal flexural stresses. It is easily seen this is true whether the struc-
ture is loaded with one or more trucks.
Furthermore, in the finite element analysis of the bridge of Fig. 3.4,
the maximum stresses due to transverse flexure in the deck of the bridge
resulting from the load of 100 kips, applied either concentrically or ec-
centrically, were of the order of 2 260 psi. Because these loads were
applied directly over the beam webs, primary bending of the upper deck

100 KIPS


- - -
- -
f, --\\ -CH-
5 0.500 ; -..-IPPRCjXIMATE METH0IJ’ 1

20.375 r
5 4 3 2 1

Fig. 3.6 Transverse deflection of the bridge cross section shown in Fig. 3.4 due to a single
eccentrically-applied load of 100 kips as predicted by the approximate and finite
element methods.

was avoided. Hence, one can conclude that the secondary moments in-
duced by truck design wheel loads in the decks of bridges of this type
are of a minor nature under usual conditions and therefore can be ig-
nored. The design of the deck can safely be made under usual conditions
using the elastic methods of analysis described in Chapter 1 disregarding
the relatively minor stresses resulting from transverse frame action.
An approximate longitudinal elastic flexural analysis for a box-girder
bridge, which includes the effect of torsional rotation but which ignores
the effect of transverse bending, consists of using the principle of super-
position and combining the effects of the vertical deflection due to the
load to that of the rotation due to the torsional moment. The procedure
consists of computing the vertical deflection of the superstructure with
due regard to the restraint of the supports, assuming the transverse stiff-
ness to be infinite and using the section properties of the gross section.
Secondly, the effect of the rotation due to the torsional moment is de- ’

termined based upon the torsional constant of the gross section, again
neglecting the effects of transverse flexural deflection. Combining the
results of these two steps results in the approximate deflected shape of
the structure due to the loading condition under study. The flexural
stresses are proportional to the deflection of the structure. Hence, the
approximate (and conservative) values of the longitudinal flexural stresses
are easily determined.
The torsional constant of a section must be determined in order to
evaluate the effect of torsional stiffness. This can be done using the
membrane analogy (Ref. 39) or by solving the equations of equilibrium
(Ref. 40).
For the section shown in Fig. 3.7(a), shear flows exist as shown in
Fig. 3.8 due to the torsional moment MT. The equations for the web
shear flows at sections B-B’ through E-E’ are:

4ti = 42 - 41 (3.1)
4i = 43 - 42 (3.2)
4” = 43 - 4.1 (3.3)

49 = 4, - 45 (3.4)

Shear flow is equal to the product of the shear stress and the wall thick-
ness of the element under consideration.
The relationship for the torsional couple is:

24,A, + 24?A2 + 2&A, + 24,A, + 245& = MI (3.5)

in which the notation is as defined in Fig. 3.7(a), the areas are as defined
in Fig. 3.7(b) and MT is the torsional moment applied to the member.
The relationships between the shear flows and the rate of twist 8,
using the same notation from Fig. 3.7(a) are:

,,[~+++JL]-4.$=2~&~ (3.6)

4.*+4&+-$ 1 -4i +$ = 2 G.426, (3.7)

4i -
H23 +43
[ -
ttn 1 - 4 f$ * = 2 GA3B (3.8)

Fig. 3.7 Cross section of a “closed” section used to illustrate the computation of torsional
shear flow and the torsional constant.

fh4 +f14
T4 1 - 49 ++ = 2 GA40 (3.9)

4fJ 1 = 2 GA,0 (3.10)

It should be recognized that the terms x , k , etc. in Eqs. 3.6

through 3.10 are equal to: fh, Lll

’ ds (3.11)
s 0 t -

for these members of constant thickness. For sections which have com-
ponents of variable thickness, such as haunched slabs, one must solve
Eq. 3.11 for each component of each cell. \

A 7 B’ 7 c’ -q- D’ 4 E’T F

Fig. 3.8 Diagram of shear flows for the cross section of Fig. 3.7 when subjected to a torsional
moment M T .

The relationship between the rate of twist and the applied moment is:

(j=EL (3.12)

in which 8 is the rate of twist resulting from moment M T, J is the tor-

sional constant and G is the shear modulus. This relationship can be
G,j = 9 (3.13)
For a moment equal to unity,

GO = + (3.14)

Hence, in equations 3.6 through 3.10 the terms on the right side can be
t a k e n e q u a l t2A1
o -- - 2Az 9. . . . . .-and
2A5 in equation 3.5 the
term on the right side can be taken equal to one. Solving these equations
will give the values of 41 through & and J.
The unit shear stresses due to torsion can be found by dividing the shear
flow for any element by its thickness. This is:

VI = f (3.15)

in which VI is the torsional shear stress.

Considering the box girder section shown in Fig. 3.4, and using the
principles given above, the torsional constant will be found to be of the
order of 732 ftS. The moment of inertia of the gross section is 5,663,boo

inr. From these the vertical, rotational and total deflections of the webs
one through five for a vertical load of 100 kips applied with an eccentric-
ity of seven feet can be determined. The vertical deflection is computed

6 = -- PL3 - 100 x 1203 x 1728 = 0.366 in. (3.16)

48EI 48 x 3000 x 5,663,OOO
The angle of twist at midspan is:

8- Md- - 50 x 7 x 60’ = 0.0001660 (3.17)

2GJ 732 x 1200 x 144

The resulting deflections for beams one through five are summarized in
Table 3.1 together with the deflections obtained from finite element
analysis. In each of the analyses the values for elastic modulus and
shear modulus were taken as 3000 and 1200 kips per square inch respec-

Vertical Total from a Finite
Deflection Deflection Deffection Element Analysis
6 60 61 6 F.E.
BEAM NO. fin ) (in.) (if?.) (in.)

1 -0.366 -0.028 -0.394 -0.362

2 -0.366 -0.014 -0.380 -0.378
3 -0.366 0.000 -0.366 -0.333
4 -0.366 +0.014 -0.352 -0.308
5 -0.366 +0.028 -0.338 -0.297

An examination of Table 3.1 will reveal the approximate method of

analysis predicts a slightly larger deflection for all five beams. The ratios
between the predicted deflections for the average beam (beam 3) and the
maximum and minimum deflections with each method of analysis is in
remarkably close agreement.
The AASHTO Specifications provide an empirical value of S/7.0 for
the portion of a wheel load that is to be distributed to an interior girder
of a box-girder bridge. In addition, these specifications provide that an
exterior girder of a box-girder bridge is to be designed for a live load
wheel distribution equal to Wd7.0 in which We is defined as the top slab
width measured from the midpoint between girders to the outside edge
of the slab. It should be obvious that these empirical values may or may’
6 6 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

not reasonably reflect the actual loads which are imposed upon a bridge
structure. A rational analysis, as described above, gives the designer a
good understanding of the structural behavior that can be expected from
a given set of conditions. The empirical distribution factors may be safe
(perhaps unduly safe) but do not necessarily reflect the true behavior of
the structure.
A comprehensive study of a model of a reinforced concrete box-girder
bridge has been reported (Ref. 25, 26). The tests confirm the adequacy
of usual flexural theory for the design of bridges of this type. This con-
clusion can reasonably be extended to include bridges of prestressed

3.3 Longitudinal Flexural Design

From the considerations of torsional stiffness explained above, together

with the provisions of the AASHTO Specification relative to number
and positioning of design traffic lanes, which were explained in Chapter
1, it should be apparent that the maximum longitudinal flexural stresses
occur in the box-girder bridge of usual proportions when the maximum
number of design traffic lanes is applied to the structure. This may not
be true in the case of a girder bridge. Because the eccentricity of the live
load is normally small when the maximum number of design tra!Xc lanes
is applied to the structure, the effect of the torsional moment on the
maximum longitudinal stresses should not be expected to be significant
and hence can normally be neglected without incurring serious consequ-
ences. Maximum torsional stresses are found when the bridge is subject
to less than the maximum number of design traffic lanes.
It should also be recognized that the AASHTO Specifications, in Ar-
ticle 1.3.1, stipulates an empirical load distribution factor of S/7.0 for
concrete box-girder bridges (see Table 2.1). In the case of the
superstructure shown in Fig. 3.4, the use of the requirements of the
AASHTO Specifications would result in the structure being designed
for 2.5 design lanes (five wheel loads) in spite of the fact Article 1.2.6 of
the AASHTO Specification stipulates fractional design lanes shall not
be used. Both finite element and approximate elastic methods of analysis
reveal that the distribution factor of 97.0 is conservative in the case of
this example. Designing for five wheel loads is not rational for a struc-
ture which behaves as this one does.
Hence, it is recommended that the longitudinal analysis of box-girder
bridge superstructures of usual proportions be analyzed using the con-
ventional flexural theory based upon the gross section properties of the

section. The cross section should not be divided into fictitious “in-
terior” and “exterior” girders. The effect of live load eccentricity
should be included, using the approximate method explained herein,
when it is of significant magnitude.
As stated previously, bridge superstructures which are relatively wide
in comparison to their span, should be designed with consideration being
given to transverse flexibility and its effect on the distribution of live
load in the longitudinal direction.
The question of the effect of shear lag on the accuracy of using the
gross section properties in the analysis of a box-girder bridge is some-
times raised. Analytical studies have shown this phenomenon does not
significantly affect box-girder bridges, nor even segmental bridge
superstructures of usual proportions (Ref. 41 and 42). The use of very
thin deck slabs in box-girder and tubular-girder bridges could conceiva-
bly result in cases where shear lag should be taken into account but with
current construction methods and materials, it is doubtful if designs of
such proportions would be feasible.

3.4 Decks for Box-Girder Bridges

The great majority of box-girder bridges which have been constructed in

the United States have bridge decks which were designed with the em-
pirical relationships of the AASHTO Specifications. The live load mo-
ments induced in the webs of the box girders have generally been ig-
nored. This practice has apparently yielded satisfactory results although
it is conservative with respect to the live load moments in the slabs
and unconservative with respect to the live load moments in the webs.
Because of the great torsional stiffness of the box-girder section and
the large flexural stiffness of the transverse “frame” formed by the top
and bottom slabs in combination with the several webs of the typical box
section, moments induced in the deck due to transverse torsional or
flexural deformation due to live loads are generally very small and can
be ignored. Hence, the live load moments induced in the structure under
normal conditions can safely be analyzed using an elastic analysis as-
suming no differential deflection exists between the webs due to the live
load. That is to say it is generally safe to ignore the secondary moments
due to transverse deformation. Account should, however, be made for
any variations in deck depth as well as for the elastic restraints induced
by the webs.
For superstructures which are unusually wide in comparison to their
span or for those which are relatively shallow with respect to their 1
68 ( H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

width, transverse flexibility becomes important and approaches the be-

havior observed in slab bridges. Special consideration due to this effect
should be included in the transverse flexural design of bridges of these
Torsional stresses, due to the eccentricity of dead and live loads, are
rarely a problem in box-girder bridges due to the high torsional capacity
of the typical box-girder bridge cross section. In spite of this, torsional
stresses should be computed and taken into consideration when propor-
tioning shear reinforcing. Torsional stresses, due to bridges being con-
structed with built-in horizontal curvature, require special evaluation
(See Section 5.3).


X-- ----- - - - . - X

Fig. 3.9 Cross section of an “open” section used to illustrate the computations of fleyural
shear stresses.

3.5 Shear Distribution

The distribution of flexural shear stresses in “open” sections, such as

rectangular, T-shaped or I-shaped beams, is computed from the well
known relationship:

~ah!!QL= v yt ybds (3.18)

Ib, I b, sya

-------THIS AXIS

Fig. 3.10 Cross section of a closed section which is symmetrical about a vertical axis.

in which the terms are shown in Fig. 3.9 and are defined:
va is the unit shear stress at fiber a.
V is the shear force acting at the section under consideration.
Qa is equal to the moment of the area ofthe cross section between yt and
ya with respect to the centroidal axis of the section.
I is the moment of inertia of the cross section with respect to the
centroidal axis.
ba is the width of the section at ya above the centroidal axis.
by is the width of the section at y above the centroidal axis.
The locations of points of zero shear stress are known in open sections.
This is also true for the case ofclosed sections which are symmetrical about
a vertical axis as shown in Fig. 3.10.
In the case of a closed section that has a single cell and which is not
symmetrical about a vertical axis, the location of the points of zero shear
are not known and a special method of analysis is required to determine
the distribution of shear stresses. The special method must also be used
for sections composed of several cells, as shown in Fig. 3.7. This too is
because the points of zero shear are not known (Ref. 43). This method
of analysis is similar to that used in the analysis of a multi-celled section
for torsional stresses.
The procedure consists of assuming slits or cuts made in each of the
cells of the multi-celled section as shown in Fig. 3. Il. The existence of
the cuts results in the section being open and hence the distribution of
shear stresses can be determined by using Eq. 3.18. If the locations of
the assumed cuts do not coincide with the locations of zero shear stress,
a shear deflection will exist in the section at the location of each cut. A
shear flow, similar to that resulting from torsion, must exist to nullify
these deflections. A different shear flow is required at each cell to re-
store the deflected shape of the section. It can be shown that for a sec-


Fig 3.11 Cross section of a multi-cell section with slits assumed in the bottom of each cell in
order to render it an “open” section for purposes of shear stress analysis. \

tion having n cells, n equations of consistent deformation can be writ-

ten. The solution of these equations results in the determination of the
shear flows in each cell. The addition of the shear flows to the shears for
the open section result in the actual shear stresses for the loading condi-
tion. The equations of consistent deformation for a cell j of a multi-cell
section as shown in Fig. 3.11 takes the form:
in which
@, Z$ and @ are the shear flows for cells i, j and k respectively, that are
required to nullify the deflections at the assumed cuts in cells i, j and k.



’ ds (3.22)

in which q. is the shear flow at any point in cell j with the section being
open (assumed slits in each cell).
For cells composed of elements of constant thickness the terms &i and
& are equal to the length of the element ji and element jk divided by
their thicknesses respectively. The term sii is equal to the length of each
element forming cell j, divided by its thickness. The term sjo is easily
solved because ds / t is a constant for each element of the cell and q.
i 7476 - ls?fl -1416

Fig. 3.12 Shear flow for the cross section of Fig. 3.11 if cuts exist in the bottom of each cdl.

--a- -111 111 361-

250 1 222 1 I 250 361 I

361, ill- ,111 -361

Fig. 3.13 Shear flow required to nullify the shear deflections for the cut section.

normally varies linearly or parabolically along the length of each of the

The procedure can best be described with an example. Consider the
typical box-girder bridge cross section shown in Fig. 3.4. The moment of
inertia of the cross section is 5,663,OOO in4. The section is subject to a
vertical shear force of 600 kips. Assuming cuts in the bottom slab at the
center of each span, the shear flow for the open section is as shown in
Fig. 3.12. Solution of the equations of consistent deformation reveals the
shear flows to be as shown in Fig. 3.13. The actual distribution of shear
stress is equal to the sum of the values given in Figs. 3.12 and 3.13 and
are as shown in Fig. 3.14. Note that the shear flows in the first interior
webs are 1.17 times that in the exterior webs.


Fig. 3.14 Actual distribution of shear stress.

B O X - G I R D E R B R I D G E S 17 3

3.6 Construction Details

Depth-to-span ratios for post-tensioned box-girder bridges have fre-

quently been of the order of 1 to 25. They have, in some instances, been
constructed with depth-to-span ratios as low as 1 to 16 and as high as 1
to 27.8. Although one would normally expect continuous box-girder
bridge spans to be somewhat more slender than simple bridge spans, a
. review of existing box-girder bridges reveals the depth-to-span ratios to
be very similar for continuous and simple span structures (Ref. 44).
The web spacings for box girder bridges in the recent past have been
of the order of seven to nine feet. Although the smaller web spacings
have resulted in lower quantities of concrete and perhaps in some cases
less transverse reinforcing steel, the additional labor required for form-
ing costs associated with the smaIler web spacings have resulted in the
larger web spacings being used.
Minimum slab thicknesses for the slabs of box-girder bridges are
specified in the AASHTO Specifications, Section 1.6.24. The minimum
thickness specified for a top slab is the greater of the clear span between
webs divided by 16 or 6 inches. For the bottom slab, the minimum
thickness is specified as 5.5 inches or the clear span between webs di-
vided by 16, whichever is greater. It is frequently necessary to thicken
the bottom slab in areas of negative moment to 10 or 15 inches in order
to accommodate the flexural compression stresses which occur in the
bottom slab in these areas. For exceptionally long spans, even thicker
bottom slabs may be required in the areas of negative moment.
Fillets of the order of four inches by four inches have commonly been
provided between the slabs and the webs of box-girder bridges (see Fig.
3.4). Current practice in California gives the contractor the option of not
providing fillets between the webs and bottom slab if he so desires. The
L slabs, either top or bottom, have been traditionally made of constant
depth in box girder construction in the United States. This is probably
I the result of the facts that no advantage exists in haunching the slabs
when they are designed in accordance with the empirical coefficients of
the AASHTO Specification and forming costs are greater for slabs of
variable depth.
Box-girder bridges post tensioned with all the tendons being placed in
the webs, as has been the practice in California, are normally detailed
with webs that are twelve inches thick. Experience has shown that this
width is necessary except where the prestressing force is relatively small
in which case webs of 8 to 10 in. can be used. It is recommended the
bridge designer make a rough layout of the tendons for a typical interior
and exterior girder at the time the bridge design is being made as a ’


Effective Jacking Dimension’ Dimension”
@ 0.6Of’s @ 0.75f’s X (inches) H (inches)

960 1200 4.5 29

1440 1800 6.0 44

2160 2700 8.5 58

2640 3300 11.0 72

* = D + A in Figs. 3.15 & 3.16

** = Height in Fig. 3.17
TABLE 3 . 3

Max. Number of Min. Prestressing Force
YZ”C#J Grade 270 Duct Per Tendon (Kips)
Strands Per Dia.
Tendon fin) @ 0.60f ‘s @ 0.70f ‘s @ 0.75f ‘s’ @ 0.85 f ‘s”

12 -2% 297 347 372 396

18 3 446 520 558 595
24 3!4 595 694 743 793
31 4!4 768 896 960 1024

*Maximum jacking stress permitted by Standard Specifications of California Department of

**Maximum jacking stress permitted by AASHTO Specifications.

means of insuring the adequacy of the width of the webs for tendon and
concrete placing as well as confirming the required eccentricity of the
prestressing force can be achieved at the critical points.
An important item to be checked (See Fig. 3.15) is that adequate
space is provided for both the pier or bent cap reinforcing and the lon-
gitudinal tendons at points of negative moment. The necessity of con-
firming the adequacy of space also applies in areas of positive moment
where the longitudinal tendons must not conflict with the space occupied
by the bottom deck (See Fig. 3.16) and intermediate diaphragm reinforc-
ing (if any). As an aid in accomplishing this, the data in Tables 3.2 and
3.3 are presented. These data are to be used in conjunction with the
details shown in Fig. 3.15 and 3.16.


Fig. 3.15 Typical joint detail of top deck and web in a box girder bridge.


Fig. 3.16 Typical joint detail of bottom deck and web in a box girder bridge.

Fig. 3.17 Anchorage space requirements at the end of a typical box-girder bridge.

End diaphragms from 18 to 36 inches in thickness are normally pro-

vided at the abutment ends of box-girder spans. The end diaphragms
serve the usual purpose of an end diaphragm but also serve as an area in
which the post-tensioning tendons flare to their anchorages (Fig. 3.17).
The thicker end diaphragms are normally used for bridges with greater
Although the use of g-inch thick intermediate diaphragms has been
common in the past with this mode of construction, as well as currently
being required by Section 1.6.24 (F) of the AASHTO Specification, as
previously explained, it is expected this practice will diminish.
In applications where the conditions of the jobsite permit, multi-span
cast-in-place box-girder bridges should be provided with end spans which
are of the order of 80% of the length of the interior spans. This ratio of end
span length to length of interior spans results in a favorable distribution of
moments along the structure.
In simple span box-girder bridges, the tendons are generally placed in
the structure in such a manner that the center of gravity of the prestress-
ing steel, and hence the prestressing force, lies on a second-degree
parabolic trajectory. The center of gravity of the tendons may or may
not have a nominal eccentricity at the ends of the structure. This is
illustrated in Fig. 3.18. Trajectories formed of parabolic curves com-
B O X - G I R D E R B R I D G E S (7 7

Fig. 3.18 Typical second degree parabolic tendon trajectory for a simple span bridge.



Fig. 3.19 Tendon trajectory composed of two second degree parabolic curves connected by a

pounded with tangents, such as shown in Fig. 3.19 may have merit in
some instances because the straight portions of the tendons are more
easily tied in position during construction and the curvature of the ten-
dons can be confined to the areas where it is most beneficial in reducing
shear stresses on the concrete. The friction loss during stressing will be
less for the trajectory of Fig. 3.18 than for that of Fig. 3.19.
In continuous structures, the tendons are generally placed in such a
manner that the center of gravity of the prestressing steel is on a trajec-

Fig. 3.20 Typical tendon trajectory for a continuous box-girder bridge.

78 ) H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

\ \

Xl X3 x3

TAN al = 511%
TAN a2 = TAN a, = Z!Y,/X, = 531x3

Fig. 3.21 Tendon trajectory geometry for an end span of a continuous box-girder bridge.


TAN a, = TAN a2 = a,/~, = a,/ X2
TAN a3 = TAN a4 = 2y3/x3 = fi?yr/Xa

Fig. 3.22 Tendon trajectory geometry for an intetior span of a continuous box-girder bridge.



Fig. 3.23 Tendon layout designed to reduce shear stresses at the interior support

tory formed of compounded second-degree parabolas. This is illustrated

in Fig. 3.20 from which it will be seen that the trajectories in the end
spans are formed of three parabolic curves while that in the interior span
is composed of four parabolic curves. The geometry of the trajectories
must satisfy the requirements shown in Figs. 3.21 and 3.22.*
Although it is somewhat less efficient from a flexural standpoint, the
tendon layout shown in Fig. 3.23 is more efficient in reducing shear
stresses on the concrete section than the trajectory of Fig. 3.20.
Different jurisdictions have different clearance requirements for post-
tensioning ducts. The requirements of the California Department of
Transportation which are applicable to precast post-tensioned girders
are illustrated in Fig. 3.24. The clearance requirements for post-
tensioned ducts in cast-in-place box-girder construction as specified by
the California Department of Transportation are shown in Fig. 3.25.

*The designer must evaluate the friction losses and effects of anchorage deformation which
occur during stressing continuous tendons of this type in order to ascertain the minimum \
amount of prestressing that will satisfy the design criteria (Ref. 45).
8 0 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

The pleasant appearance of simple and continuous span box-girder

bridges is illustrated in Figs. 3.26 and 3.27. Note the fact that the bent
cap is hidden within the depth of the bridge superstructure in Fig. 3.26.
The ease with which horizontal and vertical curvature as well as
superelevation is accommodated with this mode of construction is illus-
trated in Fig. 3.27.

8 8 8

Fig. 3.24 Clearance requirements for post-tensioned ducts in precast girders according to the
requirements of the California Department of Transportation.



I 1

DUCTS 4” TO 4%” O.D. DUCTS OVER 4%” O.D.

Fig. 3.25 Clearance requirements for post-tensioned ducts in cast-in-place box-girder

bridges according to the standards of the California Department of Transportation.
Fig. 3.26 Typical box-girder bridge with bent cap within the superstructure. (Courtesy of
California Department of Transportation.)

Fig. 3.27 Mission Valley Viaduct in San Diego, California. (Courtesy of California Department
of Transportation.)
4 Segmental

4.1 Introduction

The term “segmental box-girder bridge” or “segmental bridge” has been

used in this book to describe special forms of box-girder bridges. The
special forms are characterized by three basic cross sections which are
illustrated in Figs. 4.1 through 4.3. The fundamental difference between
the cross sections of the segmental bridges and box-girder bridges is the

I Fig. 4.1 Typical cross section of a segmental bridge.

84 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

& Fig. 4.2 Typical cross section of a segmental bridge.

Fig. 4.3 Typical cross section of a segmental bridge.

relatively few webs incorporated in the segmental bridge cross sections.

Another difference is that in segmental bridges, as in the case shown in Fig.
4.3, the bottom slab may not be completely continuous between the most
exterior webs. Additionally, segmental-girder bridges generally employ
deck slabs of variable depth and are constructed without intermediate
diaphragms except at locations of hinges. Hinges are sometimes provided
within a span in order to accommodate length changes resulting from
concrete shrinkage, creep and temperature variations. The typical cross
sections used in’segmental girder bridges have evolved in Europe where
they have been optimized with a view toward reducing their dead load. The
special importance of superstructure dead load with long span bridges is
well known.
Torsional stiffness, as in the case of box-girder bridges, is important in
the manner in which segmental bridges behave structurally. With the.
exception of superstructures formed of two or more tubular girders con-
nected with a common slab, the structural behavior of a segmental-girder
bridge is basically the same as that of a box-girder bridge. The special
considerations for the multi-tubular girder bridge with connecting slabs
between the individual girders is limited to the special analysis required
for the transverse distribution of moments discussed in Section 4.4.
Because of the unique cross sections and erection techniques used

with this mode of bridge construction, special design considerations

must be employed in their proportioning and analysis.

4.2 Longitudinal Flexural Analysis

The fundamental principles of structural analysis used in the design of

bridges of other types are used for the determination of the maxima and
minima moments and shears which can occur along the length of seg-
mental box-girder bridges. However, certain types of flexural analyses
not required in the design of bridges of other types are required in the
design of segmental bridges that are erected in cantilever or with other
special methods of erection. These include the dead load analysis for the
statically determinate cantilevers, the determination of the secondary
moments which occur at the various stages of construction as a result of
the continuity tendons being installed and prestressed, and the analysis
of the redistribution of moments (see Section 4.3) which occurs in the
structure as a result of creep, shrinkage and relaxation. The effects of
superimposed dead and live loads are determined as for any other struc-
The cantilever erection technique is illustrated in Fig. 4.4. It will be
seen that the procedure consists of first setting the pier segment on top
of the completed pier after which span segments are placed first on one
side and then on the other. The segments may be cast-in-place or
may be precast. In either case, the segments are prestressed to the pre-
viously constructed work, after they are placed.
It should be noted that some means of accommodating the unbalanced
moment must be provided with a superstructure erection technique such
as this one. The unbalanced moment can be accommodated by securely
attaching the pier segment to the pier, either permanently or temporarily
during construction, if the pier itself is sufficiently strong to accommo-
date the unbalanced moment. A discussion of special considerations for
moment resisting piers to be used in bridges erected in cantilever will be
found in Section 5.6. In some cases temporary bracing struts or tempor-
ary ties are used to accommodate the unbalanced moment. These
methods are all illustrated in Fig. 4.5 and 4.6. Temporary struts and ties
can be relatively costly. The removal of a temporary tie or strut results
in the application of a load to the structure. The load in ties and struts
can be quite large and must be considered in the stress analysis of the
structure when they are used. It should be recognized that the founda-
tions must be able to withstand the unbalanced moment if the temporary
struts or ties are connected to them. Specially designed erection gantries \
6 6 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S


tr IYr?


8 Tl!?




Fig. 4.4 Cantilever erection scheme.

Fig. 4.5 Superstructure-substructure temporarily fixed for unbalanced moment.

:n I


I,,, I , , , , , , , , ,

A i

Fig. 4.6 Temporary struts for unbalanced moments.





Fig. 4.7 Unbalanced moment due to form travelers in cast-in-place segmental construction.

which are capable of erecting precast segments as well as bracing the

superstructure for unbalanced moments during erection have been used
successfully. The unbalanced moment which occurs in cast-in-place
segmental construction, due to the weight of the travelling forms, is
shown in Fig. 4.7.
The computation of the cantilever moments which occur during the
erection process is simple and straightforward. They must be computed
for the location of every joint between the segments because the stresses in
the concrete due to dead load and the initial prestressing forces must be
checked at each joint at each step in the erection procedure. A diagram
illustrating the cantilever moments which occur during erection is given in
Fig. 4.8.
The overall erection of a typical bridge superstructure consisting of
five spans is shown in Fig. 4.9. This particular superstructure is sup
ported by hinged bearings at the top of each pier and abutment. The
superstructure erection commences with the erection of the cantilevers
which extend from each side of the first pier. The second step consists
of constructing the end portion of the first span. When the end portion is
completed, it is connected to the cantilevered superstructure con-
structed in the first step by the placing of the closure joint. The first
span is completed by the installation and stressing of the “continuity
tendons” which extend between the two portions of the span. Because
there are only two simple supports, the stressing of the first “continuity
tendons” does not create secondary moments in the structure. The’erec-


tlj10~9~8~7~6~5j4~3~2jl jl 12j3j415j61718j9110111
I )
Fig. 4.8 Moments during erection of a cantilever.

tion progresses with the construction of the cantilevers which extend

outwards on each side of the second pier (Fig. 4.9~). When the cantilev-
ers near midspan of the second span are joined together and the closure
joint has been placed and cured, the continuity tendons joining the two
cantilevered portions of span two are inserted and stressed. Because the
structure is continuous over three supports at the time the second group
of continuity tendons is stressed, secondary moments, which have a dis-
tribution as shown in Fig. 4.10a, are created. Repeating this procedure,
the construction continues until span five is completed and rendered
continuous to the previously erected structure. The secondary moments
resulting from the stressing of the various groups of continuity tendons
together with their sum, are shown in Fig. 4.10. It should be noted that
an extension to the cantilever of the first portion of the third span must
be placed after continuity is established in the second span. This is the
result of the unequal span lengths.
Other erection techniques which lend themselves to precast segmental
bridges also result in secondary moments, as well as primary moments,
which must be taken into account in the superstructure design. Consider
the two span freeway overcrossing structure shown in Fig. 4.11. This
structure is erected using temporary towers together with temporary
prestressing. The erection procedure consists of erecting the three as-
semblies of precast units on top of the towers using the temporary pre- ’
stressing tendons as shown in Fig. 4.12(a). After the three assemblies
90 ) H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

155’ 210 155’






t t

I I 1 I

t t t


Fig. 4.9 Cpnstruction sequence for a five-span bridge erected in cantilever.


I / I 1


/I ! 1
I I t

I I I \ I




I -

, I I


Fig. 4.10 Secondary moments which occur during erection of the bridge shown in Fig. 4.9.

Fig. 4.11 Two-span freeway overcrossing structure.




Fig. 4.12 Erection of a freeway overcrossing structure using temporary towers and tempor-
ary prestressing.

135’~0” 135’-0” d

Fig. 4.13 Two span highway overcrossing structure. .


are erected, the two outermost assemblies are slid towards the middle
assembly until they are in contact as shown in Fig. 4.12(b). It should be
recognized that all of the precast units fit together perfectly because they
are precast one against the other in the same order as they are to be
assembled in the structure. This procedure also results in the outer as-
semblies of precast units fitting the inner assembly of precast units per-
fectly and eliminates the need for cast-in-place joints between the as-
semblies. In addition, hydraulic jacks are provided between the as-
semblies of segments and the temporary erection towers. (See Fig.
4.12a). The hydraulic jacks allow realignment of the assemblies during
erection in the event it is necessary to correct for settlement of the tem-
porary towers or make other minor adjustments. After the assemblies
have been slid together, the first stage permanent prestressing tendons,
having the shape shown in Fig. 4.13, are installed in preformed ducts in
the segments and prestressed. Because the superstructure is continuous
over six supports at the time the first stage tendons are stressed, a sec-
ondary moment as shown in Fig. 4.14 is created by the stressing. (Note:
The hydraulic jacks at the top of each set of twin towers are intercon-
nected during this operation with the result they act as a “hydraulic

Fig. 4.14 Secondary moment due to first stage prestress for the bridge of Fig. 4.11.


Fig. 4.15 Moment due to removal of the temporary erection towers shown in Fig. 4.12.

Fig. 4.16 Secondary moment due to removal of the temporary prestressing tendons.

Fig. 4.17 Secondary moment due to second stage prestressing.

hinge” and the two jacks can be considered as one point of support.)
The next step in the erection consists of removing the erection towers.
This results in a primary moment, having the shape shown in Fig. 4.15,
being imposed upon the structure. Removal of the temporary prestress-
ing tendons constitutes the next step in the procedure. This results in
the secondary moment distributed as shown in Fig. 4.16. The secondary
moment due to the second stage permanent prestressing tendons is
shown in Fig. 4.17. Installation and prestressing of the second stage ten-
dons completes the erection procedure.
From the above description of the various different moments which
occur in bridges that are erected segmentally it should be apparent that
the design of bridges of these types requires a considerable amount of
engineering effort. Because the moments are dependent upon the con-
struction methods that are to be used, the designer must specify the
construction sequence that is to be followed. Because the stresses at
each joint between the segments must be considered for each step in
construction more design effort is required for segmental bridges than
for more conventional bridge types. It should also be recognized from
the above discussion that the basic elastic analysis involves methods
familiar to all structural engineers.
The effect of live load on a segmental bridge must be found using
fundamental engineering principles together with the basic provisions of
Section 1.2 of the AASHTO Specifications. The provisions of Section
1.3 are clearly not intended to apply to bridge superstructures composed

of a single tubular girder. There are few complexities involved in the

theoretical analysis of the distribution of wheel loads to members of this
type. Consider the cross section of the single tubular girder shown in
Fig. 4.18 which is subjected to an eccentrically applied vertical load.
The eccentrically applied load can be analyzed as a concentric load and
a torsional moment. Due to the great stiffness of the section, the rota-
tion is small with the result the deflection of point A is only slightly
greater than the average vertical deflection of the section.
It can be shown that for eccentrically applied live loads that are large
in comparison to the total load, longitudinal warping stresses can be sig-
nificant (Ref. 27). For most applications where segmental bridges will be
considered for use in North America, the live loads used in design will
not be large in comparison to the total load. Hence, this effect is gener-
ally ignored and, as far as the consideration of longitudinal flexural
stresses are concerned, the loads are assumed to be applied concentri-
cally. The torsional stresses should be considered using the greatest tor-
sional moment that can be obtained with the live loads positioned as
provided in the AASHTO Specifications.
As in the case of box-girder bridges, it will be found that segmental-
girder bridges should be designed for longitudinal moments resulting L

Fig. 4.18 Cross section of a typical segmental bridge having a single tubular girder. ’

from the greatest number of design traffic lanes that can be placed upon
the structure. The design for torsional stresses should be based upon the
number of design traffic lanes which results in the greatest torsional
The usual “beam” or “flexural” theory is adequate for the analysis of
longitudinal flexural and shear stresses of most bridges which are to be
built segmentally. The effect of the transverse flexural and torsional de-
formations on the longitudinal flexural and shear stresses is small and
hence can safely be ignored. Bridge superstructures which are wide in
comparison to their span or depth, require special consideration because
of the transverse deformations which result when concentrated loads act
upon them. This is not unique to bridge superstructures which are to be
constructed segmentally but is also true for slab, girder or box-girder
bridge superstructures which are constructed using conventional
Shear lag is not normally a problem in segmental bridge superstruc-
tures of normal proportions (Ref. 41). Segmental bridge superstructures
of usual proportions, which is intended to mean relativly narrow in
width in comparison to the span length, are generally analyzed for fle-
xure using the section properties of the gross section. Experience and
analytical studies have shown this to be a reasonable assumption (Ref.
Although the torsional stiffness of the tubular girders eliminates the
need for intermediate diaphragms, diaphragms are normally needed at
the points of support. Support diaphragms are used to transfer vertical
loads as well as longitudinal and torsional moments to the substructure.
This important need should not be overlooked. In a number of small
structures constructed in France, diaphragms have been provided at the
abutments alone with the bearing details at the pier-superstructure con-
nections being such as to preclude the transfer of longitudinal bending
moments to the substructure.

4.3 Creep Redistribution of Moments

The redistribution of moment due to concrete creep constitutes an im-

portant design consideration for segmental bridge superstructures that
are erected in cantilever. This phenomenon is unique to structures
which are erected in one structural configuration which is subsequently
altered into another as is done with the cantilever erection technique.
There is no redistribution of the moments in structures which are con-
structed at one time as is the case with conventionally cast-in-place con-

Crete structures or with segmental bridges that are assembled and stres-
sed in one operation. This phenomenon is generally characterized by a
relatively small reduction in the effective negative moment at the piers
and a relatively great increase in the effective positive moment at
midspan. Effective moment is defined as that which is due to the com-
bined effects of dead load and prestressing.
The cause of the redistribution of moment can be illustrated by con-
sidering two cantilever beams as shown in Fig. 4.19. These beams are
rendered continuous when the cast-in-place joint between them is con-


Fig. 4.19 Two cantilever beams used to illustrate the cause of moment redistribution due to

strutted and the continuity tendons between the two cantilevers are in-
stalled. It should be apparent that if the two beams were not rendered
continuous, the effect of creep would be to cause vertical deflection and
a rotation of the ends of the beams with the passing of time. Because
this rotation is prevented by the provision of continuity, a positive mo-
ment is created near midspan. The creep-caused positive moment is of
significant magnitude and is accompanied by a reduction of the negative
moment at the supports which is normally of negligible magnitude.
The existence of creep-induced moment redistribution was not recog-,
nized in the early applications of segmental bridges. The result was that
98 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

some months after being put into service, cracks developed in the tensile
flanges of some of these bridges in the areas of high positive moment.
Although the bridges were not considered to be dangerous as far as their
ultimate strength was concerned, it was felt their durability might be
adversely affected by the existence of the cracks. For this reason, addi-
tional positive moment prestressing was introduced to close the cracks.
An early discussion of this phenomenon appeared in a paper by Jean
Muller (Ref. 46) in 1967. In this discusssion Muller has shown an exam-
ple of a structure where it is estimated (by approximate calculations)
that the effect of creep would reduce the effective negative dead load
moment by 3% and increase the effective positive dead load moment by
20%. A more sophisticated method of evaluating the effect of this
phenomenon is to use a numerical integration process in which concrete
strength, modulus of elasticity, shrinkage and creep as well as relaxation
of the prestressing steel are all treated as time functions using the
methods first proposed by Subcommittee 5 of AC1 Committee 435 (Ref.
47). The results of such an analysis, together with the effect of superim-
posed dead loads, can be used to determine the time-dependent rotations
at various sections along a span. The rotations can then be used to de-
termine incremental changes in the distribution of moments throughout
the bridge structure as well as the changes in stress which result there-
from. If one makes these computations using numerical integration with
a small time interval, and carries out the computations for a total time
period (for the youngest concrete) of the order of 1,000 days, substan-
tially limiting results will be obtained. An electronic computer is re-
quired for these computations due to the large number of variables in-
volved in the analysis.
An approximate method which has been used in France to provide for
the positive moment which results from the redistribution is to simply
provide a residual compression of at least 225 psi under the effects of
dead plus live plus impact loads in areas of positive moment. Because of
the smaller design live loads used in this country, this latter method may
not be conservative in the United States.
As an alternative to attempting to compute or estimate the magnitude
of the redistribution of moments resulting from the creep of structures
erected in cantilever, provision for future additional positive moment
prestressing can be made during construction. This is accomplished by
providing extra ducts for ‘prestressing tendons in the positive moment
areas during the original construction. In this manner, additional posi-
tive moment prestressing tenddns can be added to the structure anytime
after its completion in the event distress is subsequently detected in the
areas of positive moment. I

A method of computing the creep-induced positive moment which will

develop at midspan of two cantilever beams which have been rendered
continuous has been proposed by M. Thenoz (Ref. 48). The relationship
resulting from this method gives the following expression for the positive
moment due to the redistribution due to creep:

M = _ Wl @ [ 1 - r (tl - t0) ]
02 [l + Q]

in which
01 is the rotation in radians at the end of the cantilever due to the
combined effects of dead load and prestressing which the can-
tilever would undergo if not rendered continuous, with the elastic
modulus being taken as unity. (See Fig. 4.20 for sign convention).
02 is the rotation at the end of the cantilever due to a unit moment
applied at the end of the cantilever, with the elastic modulus
being taken as unity. Units are radians (per unit moment).
to is the time in days at which the dead load and prestressing are
t1 is the time in days at which continuity is established.
a is a coefficient dependent upon the concrete quality, creep
maturity (age at loading) and theoretical thickness of the con-
crete as well as the creep-time coefftcient.

0 = KR r(tx - t 0) (4.2)

Ku and r(r) are from the Appendix I of the French prestressed concrete
code which appears as Appendix A in this book.
The numerator of Eq. 4.1 is proportional to the amount of rotation the
end of the cantilever would undergo under the action of dead load and

Fig. 4.20 Sign convention for creep redistribution calculation method proposed by Thenoz. ’

prestressing subsequent to the time continuity was established [from time

tl to t -1. The denominator is proportional to the total amount of rotation
the end of the cantilever would undergo (elastic and time dependent) under
the effect of a unit moment applied at time tl. Hence, the numerical
integration procedure for the computation of time dependent deflections
and rotations (Ref. 47) could easily be adapted to determine the moment
resulting from the actions of concrete creep and shrinkage together with
relaxation of the steel using the principal proposed by Thenoz.
When applying Thenoz’s method, one must recognize that there is no
provision for distribution of moment to the supports and to adjacent spans.
The method is based upon the assumption that the supports of the cantilev-
ers are fixed. The effect of distribution to adjacent spans should be consi-

4.4 Transverse Flexure

The design of superstructures incorporating a single box or tubular girder

for stresses which exist in the transverse direction, consists of determining
the adequacy of the concrete section as well as the amount of prestressing
or reinforcing steel required to resist the flexural and shear stresses which
are induced in the section by dead and live loads as well as the prestressing
and temperature effects. A second consideration is the determination of
the torsional stresses which could exist in the section due to live loads
which might be applied eccentrically with respect to the longitudinal axis of
the structure or due to horizontal curvature which is to be built into the
structure. Superstructures which are formed oftwo or more tubulargirders
connected together by a slab spanning between them require special study
because the vertical deflection and rotation of the tubular girders due to the
applied loads induce stresses in the superstructure for which provision
must be made.
The analysis of the single tubular segment for transverse bending mo-
ments consists of solving a tube-frame composed of four or more compo-
nents each of which frequently has variable depth. The determination of
the moments for dead load is relatively straightforward. The analysis for
live load and impact is somewhat more complicated due to the complica-
tion of determining the effect of a series of wheel loads. The wheel loads are
applied to an elastic plate of variable depth and great width. The plate being
one of several components of the tube-frame. The solution of this problem
is greatly facilitated by the use of charts of influence surfaces such as those
compiled by Homberg (Ref. 18).
The use of the charts in the determination of live load moments for the



Fig. 4.21 Moment diagram for a single-cell tubular girder (a) dead load of section plus live
load on cantilevers, (b) dead load of section plus live load on center span only, and
(c) for transverse post-tensioning of the upper slab, only.
1 0 2 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

cantilever overhangs is straightforward. One simply computes the volume

the footprint each of the wheel loads makes on the influence surface. The
sum of the products of each of these volumes together with the magnitude
of the corresponding wheel load is equal to the moment per unit length for
the point under study. One must use a trial and error procedure to deter-
mine the maximum value. (See Section 1.4).
In the case of interior spans, the influence surfaces are used to determine
the fixed end moments for various positions of the load. The fixed end
moments are then used in a frame analysis to determine the effect of live
and impact loads on the frame. A convenient method is to construct
influence lines for moment at the critical points in the span, based upon the
results of a series of frame analyses using fixed-end moments from the
appropriate influence surface chart.
This method of designing the structure for the effect of concentrated
wheel loads is relatively simple and straightforward. It is based upon
fundamental elastic design principles rather than upon empirical coefli-
cients. In addition, it permits one to take the effect of haunches into
The results of such an analysis are shown in Fig. 4.21. The condition of
moment within the tube-frame for maximum load on the cantilevered slab
overhangs is shown in Fig. 4.21(a). The moment diagram for maximum
positive moment for the interior deck span is shown in Fig. 4.21(b). It
should be noted that the moments in the webs where they intersect the deck
are of opposite sign under these different conditions of loading.
It is sometimes preferable to post-tension the deck transversely rather
than employ conventional reinforcing steel to resist the transverse bending
moments. When the deck is post-tensioned it undergoes an elastic shorten-
ing and moments are created in the tube-frame. Due to the effects of
concrete creep, shrinkage and the relaxation of the prestressing steel, these
moments diminish with the passing of time. Depending upon the creep
characteristics of the concrete, the moments due to this effect may eventu-
ally be as little as one-third of their initial value. The moment diagram
shown in Fig. 4.21(c) is typical of those due to elastic shortening of the
If the deck and bottom slab of a box section have different temperatures,
moments are induced in the tube-frame. The moments so induced are
similar to those shown in Fig. 4.21(c) but may be of opposite sign. Because
temperature differentials ofthis type are transient ratherthan permanent in
nature, creep of the concrete does not reduce the moments caused by them.
This phenomenon should be considered in the design of the tube-frame.
It should also be noted that reinforcing steel, that is in addition to the
reinforcing steel required for shear stresses, must be provided in the webs

4’ 6 6 it 6 6 4’
,, 1 1 V


2 6’ I 6’ 2’

t Fig. 4.22 Moment diagram for concentrated live loads on the center slab of a single-cell
tubular girder.
1 0 4 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

in order to control the flexural tensile stresses in the webs. The number and
positions of live loads which result in the greatest requirement for shear
reinforcing at a section along the length of a beam, may be different from
that which produces the maximum web flexural stresses. Because of this,
the amount of web reinforcing required for the critical combinations of
loading at any section might be less than the sum of the amounts required
for shear and flexure if acting alone. This is illustrated in Fig. 4.22. The
loading condition for maximum shear, for the cross section under consid-
eration, is shown in Fig. 4.22(a) while those for two different conditions of
web flexure are shown in Figs. 4.22(b) and 4.22(c).
Special attention must be given to the manner in which the web reinforc-
ing and deck reinforcing are anchored in the web-deck joints. The details
used must insure that the reinforcing that terminates in the joint is
adequately anchored for the use intended and that continuity of the rein-
forcing steel is provided through the joint for the transfer of positive and
negative moments which act upon the joint. This is illustrated in Fig. 4.23.
Reinforcing steel bars must not be bent around re-entrant corners as shown
in Fig. 4.24 unless properly tied, so that lateral force generated by a tensile
force in the bar cannot spa11 the concrete. ’
The dimensions selected for the web-deck joint are not always com-
pletely dictated by considerations of transverse flexure. In the case where
the longitudinal prestressing tendons may consist of a number of seven-
wire strands, the size of the joint may have to be increased to accommodate
the large number of tendons that must be concentrated in or near the joint.
This is illustrated in Fig. 4.25. This problem is not as severe in structures
utilizing a large number of short straight bar tendons because the short bar
tendons are terminated in the joints between the segments, including the
deck area, rather than being bent downward from the deck into the web to
their anchorages as is normally the case with multi-strand tendons. A
typical cross section for a tubular girder with bar tendons is shown in Fig.
As mentioned above, the bridge designer has the option of reinforcing
the bridge deck with ordinary reinforcing steel or with prestressed rein-
forcement. It is impossible to give specific limitations as to which of the
two should be used because there are so many variables which should be
considered. If designed without tensile stresses in the concrete under
service loads, prestressed decks would be expected to be virtually crack-
free and more durable than those utilizing non-prestressed reinforcement.
On the other hand, segmental girder bridges generally have many trans-
verse construction joints in the structure and hence cannot be assumed to
be crack-free. Some engineers believe the existence of the many construc-
tion joints necessitates the use of a waterproof membrane on the deck of all

Fig. 4.23 Properly detailed reinforcing steel for a web-slab joint.


Fig. 4.24 Reinforcing steel improperly bent around a re-entrant corner.

bridges constructed segmentally. One might conclude that the presence of

a waterproof membrane renders the existence of fine flexural cracks in a
non-prestressed deck unimportant. The relative economics of prestressed
and non-prestressed reinforcement must be considered for each structure
taking into account the parameters of the specific project and the economic
conditions prevailing at the time the project is undertaken.
As previously stated, superstructures composed of two or more tubular
girders connected together by a common slab as shown in Fig. 4.27, are not
normally provided with intermediate diaphragms. Flexurally and torsion-
ally stiff diaphragms are, however, almost invariably provided at the points
of support for such superstructures.
Under the action of concentrated live loads, eccentrically applied be-
tween the supports, a structure of this type deforms as shown in Fig. 4.28.
The deformation of the structure under this loading is characterized by
vertical deflections and rotations of the individual tubular girders. These
deformations induce secondary moments in the connecting slab which may
or ‘may not be additive to primary moments which result from live loads!


Fig. 4.25 Web-deck joint showing longitudinal and transverse reinforcement.


Fig. 4.26 Details of prestressing in the deck of a bridge prestressed with bar tendons.

1 2 3 4

---Jzr -El-

Fig. 4.27 Bridge superstructure composed of two tubular girders connected together with a
common slab and without intermediate diaphragms.

applied directly to the connecting slab. The secondary moments in the

connecting slab can be of significant magnitude near midspan ofthe struc-
ture and must be considered in the design of bridge decks ofthis type. Near
the supports of the superstructure the secondary moments normally dis-
appear as a result of the vertical and rotational restraints afforded by
the supports.
An approximate elastic method of analysis for the transverse distribu-
tion of the loads in bridge superstructures consisting of two or more
connected tubular girders has been reported by Muller (Ref. 49). This
method of analysis utilizes the principle of superposition as well as the
method of support constants which is explained in Appendix C. The
basic principle consists of determining the primary bending moments in
the deck independently of the secondary bending moments, determining
the secondary moments independently of the primary moments and then
combining the effects of the two.
The primary bending moments are determined by analyzing the struc-

Fig. 4.28 Deformation of a bridge superstructure composed of two tubular girders connected
by a common slab under the action of an eccentrically applied load.

I 1 2 3 4

Fig. 4.29 Idealized structure for the determination of the primary moments due to the live ’
108 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

+P +P

fY -Y
,J (,!

0 0

F 1

Fig. 4.30 Symmetrical loading.

ture for live loads under the assumption that supports exist under the
webs of the structure as shown in Fig. 4.29. The fictitious supports
would prevent vertical deflection of the girders as well as their rotation.
This type of analysis is well known to all structural engineers and does
not require explanation. It will yield results that can be summarized as
influence lines for primary transverse bending in the deck at the critical
The secondary bending moments for a line load applied to the struc-
ture are determined by replacing the single line load with two combina-
tions of line loads and line moments which are equivalent to the single
line load being considered. The replacement loads and moments are

+P -P


Fig. 4.31 Asymmetrical loading.


Fig. 4.32 Eccentrically applied load.

applied directly above the webs in order to avoid primary bending. The
replacement loads and moments can be made to be in two distinct forms
which are termed symmetrical and asymmetrical. These are illustrated in
Figs. 4.30 and 4.31. As an example, the load shown in Fig. 4.32, can be
replaced by the equivalent loading conditions shown in Figs. 4.33 and
The transverse deformations for the symmetrical portion of the equi-
valent loading are shown in Fig. 4.35. Because of the symmetry, the two
girders have identical deflected shapes in the longitudinal direction. This
shape is as shown in Fig. 4.36. The reversals in curvature near the sup
ports are due to end restraints on the girders. The end restraints may be

P/4 = P/4 R, = PI4


I 0

2d 1 I
Fig. 4.33 Symmetrical component of loading shown in Fig. 4.32.
1 1 0 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

due to the flexural and torsional stiffness of the supports, due to adja-
cent spans, or both.
In the case of symmetrical loading, there are three unknowns which
vary along the span length. These are the vertical deflection (y) of the
girders, the angle of twist of the girders (0) and the moment (m) in the
connecting slab. Because of the symmetry, the connecting slab is subject
to a constant moment, as shown in Fig. 4.37 and is without a shear

R, = P/4 R, = P/4 R, = -P/4 R, = -P/4


11 ! ! ’

Fig. 4.34 Asymmetrical component of loading shown in Fig. 4.32

+P +P

Fig. 4.35 Transverse deformations for symmetrical loading.

i force. The determination of the unknowns requires the solution of three

t equations which are:
EI d4y=
dx4 ’

G-&-(J$-)= -y + m

K+Oa+b (4.5)

In the above equations, the terms are defined as follows:

E = elastic modulus of the concrete (psi).
G = shear modulus of the concrete (psi).
I = moment of inertia of each tubular girder (in”).
P = a uniformly distributed line load ofconstant magnitude extending
from support to support (plf).
w = rotation of a tubular girder (radians).
Y = uniformly distributed line moment applied to the tubular girder
from support to support (ft-lbs).
m = moment in the connecting slab (ft-lbs).
a,b = elasticity constants as defined in Appendix C for the method of
support constants.
K = flexural elasticity of the joint at each end of the connecting slab.
J = torsional constant.
For members having a non-variable torsional constant (J), Equation
4.4 becomes
GJ 3 = -y + m (4.6)


Fig. 4.36 Longitudinal deformations due to symmetrical loading of girders 1 and 2.

1 1 2 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

/ L, = 2d

m, = m,

Fig. 4.37 Shear and moment conditions for the connecting slab under symmetrical loading.

Fig. 4.38 Transverse deformation for asymmetrical loading.



m = qd

Fig. 4.39 Shear and moment conditions for the connecting slab under asymmetrical loading.

Under the asymmetrical portion of the equivalent loading, the struc-

ture deforms as shown in Fig. 4.38. In this case., there is no moment in
the connecting slab at its midspan and the slab is subjected to a constant
shear of q. The end moments for the connecting slab, as shown in Fig.
4.39, are equal to:
m = qd (4.7)

There remain three unknown variables. These are the girder deflections
(y), the girder rotations (w) and the shear force (q) in the connecting
slab. The equations for the three unknowns are:


which for a non-variable torsional constant (J) becomes

GJ $ = -y + d’q

y = -d’w + d2q (K + a - b) (4.11)
All of the terms in the above equations have been defined except for d
andd’ which are defined in Fig. 4.33.
The longitudinal deformation of the two girders under asymmetrical
loading may be as shown in Fig. 4.40. The torsional stiffness of the end
diaphragms together with the flexural stiffness of adjacent spans account
for the reversal of curvature in the curves.
A correct solution requires the restraints of the supports be taken into
account in the analysis. For symmetrical loading, both girders deflect
vertically as well as rotate equally at the supports. Therefore, they do
not induce a torsional moment in the support diaphragms. Because the
two girders have equal but opposite support torsional moments when
symmetically loaded, a circular bending will be induced in the support
diaphragms. The bending deformation of the support diaphragms due to
the torsional moments is generally small and can normally be neglected.
In the case of asymmetrical loading the vertical deflections of the girders
are in opposite directions and hence a torsional moment is induced in the
support diaphragms. The torsional rigidity of each support diaphragm is
additive to the flexural rigidity of the beams from the adjacent span, if
any exists. The deformations of the support diaphragms are generally ’
1 1 4 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S


Fig. 4.40 Longitudinal deformations due to asymmetrical loading of girder 1 and 2.

very small and can be neglected without serious errors being introduced
into the calculations. Flexural rigidity of adjacent spans or of a support,
on the other hand, have an important influence on the deflection curve
of a girder and hence should not be ignored.
The six equations, 4.3, 4.4, 4.5, 4.8, 4.9 and 4.11 can be transformed
into linear equations by assuming a particular distribution of the shear
force (q) and moment (m) in the connecting slab. It is known the dis-
tribution of the shear force (q) closely follows the shape of deflection
curve (y) of the girders. Hence, the value of the shear force (q) can be
assumed to vary parabolically from zero at the support to a maximum
value (qJ at midspan without the introduction of significant error. In this
case, the magnitude of the shear force (q,) at a distance of x from
midspan is:
qx =qcl(l i24x2) (4.12)

The angle of twist of a girder under a torsional moment (m) is equal


w = Ktm
in which KI is the torsional elasticity constant of the girder. The actual
value of Kt is maximum at midspan and is equal to zero ‘at the (torsion-

ally restrained) support. The value of KI can be assumed to be constant

along the span without the introduction of significant error. The value of
KI for a constant moment (m) along the span is
Kt’ =&.e (4.14)
while for a moment which varies parabolically from m at midspan to
zero at the supports the value of Kt is:

K;=LLZ (4.15)
48 GJ
With the simplifying assumptions described above, it can be shown
the values for the shear q and moment m in the connecting slab can be
determined for a single line load (p), uniformly distributed along the
span, by first determining the primary moments for unit loads in the
frame under the assumption fictitious supports exist under the webs of
the tubular girders as shown in Fig. 4.29. From the primary moments so
obtained, the reactions at points 1,2,3 and 4, as shown in Fig. 4.29, are
determined. The terms L1 through L5 are defined in Fig. 4.29. Using
these reactions, one computes the equivalent loads and moments on gir-
ders 1 and 2 as follows:
1’1 = R, + R, (4.16)
~1 = (Rz - R,) +-
p2 = Ra + R, (4.18)
y2 = (Rq - R,) L (4.19)

From these, the values of the corresponding symmetrical and asymmetri-

cal girder loads and moments are determined from the following:

p’, = li?+lL (4.20)

Yl ‘=Y (4.21)

I, _
- Pl - P2- (4.22)
Pl 2

!I _
- y1 + y2
- (4.23)
1 1 6 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

From these values, the.moment due to symmetrical loading m, is computed

m, = ~1 ‘Kt’ (4.24)
K + a3 + b3 + Kt”

in which K, a3 and b3 are the elasticity constants for the connecting slab.
The shear in the connecting slab due to the asymmetrical vertical loads
are determined from:
q2 = yfJ + K: (L, + L3)2 + L32 (K + a3 - b3)
4 4
in which y’ is the midspan deflection of one of the tubular girders, with due
regard to the restraint of the supports, under a uniformly distributed unit
load and y” is the midspan deflection of one of the tubular girders due to a
parabolically distributed load having a value of unity at midspan and zero at
the supports, with due regard to the elasticity of the supports. From this,
one obtains,

The shear and moment in the connecting slab due to the asymmetrical
moments are computed from:

ylr) Kt’ tL2 + Ld

? _-.
q3 = y” + Kr” (L2 + L3)2 + L32 (K + a3 - b3)
4 4

Fig. 4.41 Location of the critical section for negative moment in the connecting slab.

m 3 = q3 r

The total support secondary moments at points 2 and 3 are equal to:

m = ml + m2 + m3 (4.29)

and the total secondary shear in the connecting slab is:

q = q2 + q3

The secondary moment at the critical section (See Fig. 4.41) for negative
moment is:
m’ = ml + (q2 + q3) d (4.31)

The results of such an analysis can be plotted as influence lines as shown

in Fig. 4.42 in which it will be seen the primary and secondary moments at
the support, as well as their sum, are plotted together. A similar diagram for
the midspan moment is given in Fig. 4.43.
The influence lines shown in Figs. 4.42 and 4.43 were developed for line
loads rather than for concentrated loads and are intended for use with lane
loadings rather than with truck loadings. The lane loadings of the
AASHTO Specifications include a concentrated load in addition to the
uniform loads. The effect of the concentrated load can be included in the
analysis by increasing the magnitude of the uniform lane loading used in the
design as follows:

w’=w ( Yu-

in which w’ is the adjusted lane load in pounds per linear foot.

w is the lane load provided in the AASHTO Specification
Yu is the midspan deflection of a girder, with due regard for end
restraint, due to the AASHTO lane load.
yc is the same as yU except for the appropriate concentrated load
applied at midspan of the girder.
The lane loading generally controls for bridges of long span.
For bridges of short and moderate span, the truck loading controls the
design and the secondary moments due to girder deflection and rotation
are small and normally neglected.

Fig. 4.42 Influence line for negative moment at the critical section of the connecting slab.


I- 0.5

P l.O- \\ II


L, L, _ L == Lb>
I- -I
Fig. 4.43 Influence line for the positive moment at midspan of the connecting slab.’
S E G M E N T A L B O X - G I R D E R B R I D G E S 11 1 9

For beams which have variable moments of inertia and torsional con-
stants, a trial and error method of analysis must be used. The procedure
can be illustrated by considering the case of asymmetrical loading which
is defined by equations 4.8, 4.9 and 4.11. The specific procedure is as
1. Calculate the ordinates to the elastic curves for one of the girders
under a uniformly distributed load with due regard to end re-
2 . Assume the variation in the intensity of the shear in the connect-
ing slab q is proportional to the deflection computed in Step 1.
Designate the unknown shear force at midspan as qo.
3 . Using the variation in the shear force determined in Step 2, solve
equations 4.8 and.4.9 for the values of y and o respectively.
4 . Using the values of y and w from Step 3, solve equation 4.11 and
obtain a new value for q. at midspan and a new distribution of q
along the span.
5 . If the distribution of q obtained in Step 4 varies appreciably from
the assumed distribution, the procedure is repeated using the dis-
tribution obtained in Step 4 as the new assumed distribution of q.
It is reported the method results in rapid convergence for the effect of
the vertical load. Convergence is slower for the effect of the moments
(Ref. 49).
Influence lines of the types shown in Figs. 4.42 and 4.43 can be con-
structed from the results of finite element or other methods of structural
analysis in instances where one does not wish to employ the approxi-
mate method just described and the cost of the more sophisticated
analysis can be justified.

4.5 Proportioning the Superstructure

Before considering the proportioning of segmental bridges, a review of a

few typical bridge superstructure cross sections will be made. A review
of this type is helpful in forming a basis for future design work. Four
subject areas have been selected for consideration. These are constant
depth superstructures; variable depth superstructures; superstructures
with transverse floor beams and superstructures incorporating two or
more tubular girders.
In European practice, for reasons of economy, superstructures of
constant depth have normally been used for spans up to 200 feet. For
spans in excess of 200 feet, superstructures of variable depth have more
often been found to be more economical than constant depth structures. ’
120 ( H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

42’-0” *
~ 9’3%” ~
! J
7 ii
%8’/2” MIN.
VARIES (19’-0” MIN. 21’-0” MAX.


FT. IN.2 IN.’ Iii
8.00 8,848 6,610,000 26.73

8.00 9,663 13,350,000 36.62

10.00 10,383 22,326,OOO 45.54

Fig. 4.44 A typical midspan cross-section for the prestressed concrete segmental bridge
alternatives for the Vail pass portion of I-70 in Colorado.

On occasion, constant depth superstructures have been used for spans

considerably greater than 200 feet, generally because of esthetic consid-
erations. For optimum economy in Europe, superstructures of constant
depth that are constructed in cantilever generally have depth-to-span
ratios between one in eighteen and one in twenty. For superstructures of
variable depth, depth-to-span ratios of the order of 1 to 18 at the piers
and 1 to 40 at midspan are frequently used.
Constant depth superstructures with depth-to-span ratios approaching
1 to 30 are possible under some conditions even when the cantilever
erection technique is used. Some sacrifice in economy would be ex-
pected when the shallower depths are used. When other erection
techniques are used, such as casting the segments in place on falsework
rather than using “travelers” or using temporary falsework towers in
combination with temporary prestressing tendons in the erection of pre-
cast segments, the more slender sections should prove to be more feasi-
ble .

The lighter live loads used in United States bridge design, as com-
pared to those used in other parts of the world, should also contribute to
the feasibility of using less superstructure depth. The depth-to-span
ratios which result in optimum economy in the United States will be-
come more apparent as the use of this method becomes more common
and widespread. “Americanization” of the construction details as well
as the construction methods will undoubtedly occur. This evolution may
have a profound effect on the depth of superstructure that is most
economical for a particular structure as well as on the overall economy
of the method.
Examples of cross sections of bridge superstructures having constant
depth are those which were used on Interstate Highway I-70 over Vail
Pass in Colorado. These are illustrated in Fig. 4.44. It should be noted
that the dimensions of the upper slabs are identical for segments of all
depths as are the slopes of the webs. The principal variable dimensions
include the depth of the segments as well as the width and thickness of
the bottom slabs. The bottom slab thickness was increased near the
piers in order to accommodate the large negative moments which occur
in these areas. This method of varying the principal dimensions was
selected to facilitate the manufacture of segments having different
depths. The depth-to-span ratios used on Vail Pass, due to esthetic con-

Fig. 4.45 Typical cross-section of the Pine Valley Creek Bridge.

siderations, are generally more slender than that which would be ex-
pected by European practice to result in greatest economy.
Another example of a superstructure of constant depth is the Pine
Valley Creek Bridge which is shown in Fig. 4.45. This structure has a
main span of 450 feet and hence is in the span range that would normally
be done with a variable depth in European practice. Note that the
depth-to-span ratio for the main span of this structure is 1 to 23.7. It
should also be pointed out that the thickness of the bottom flange in the
main span of this structure varied from 6.5 feet at the piers to 10 inches
at midspan. The exceptionally great bottom flange thickness at the pier
was required as a result of using a relatively slender constant depth
superstructure with a relatively great span length.
The Saint Cloud Bridge in Paris was constructed with precast seg-
ments having the shape shown in Fig. 4.46. This structure has a
maximum span of 335 feet with a constant superstructure depth of 11.8
feet. This results in a depth-to-span ratio of 28.3. This cross section is
not the most efftcient from a structural viewpoint. The dimensions were
selected in order to achieve a desired appearance.

Fig. 4.46 St. Cloud Bridge over the Seine River in Paris, France. (Courtesy of SokiBtB
Technique pour L’Utilization de la prbcontrainte, Paris, France).
S E G M E N T A L B O X - G I R D E R B R I D G E S )1 2 3

Superstructures with variable depth can be designed with a wide com-

bination of depth-to-span ratios at pier and midspan. The intrados can be
formed of straight chords or of circular curves. Surprisingly, with seg-
mental construction parabolic curves are not desirable due to construc-
tion considerations. Spans in excess of 750 feet have been constructed
segmentally with cast-in-place superstructures of varying depth.
An interesting example of a bridge constructed with varying depth is
the Ok-on Bridge which is shown in Fig. 4.47. This structure has 26
central spans of 259 feet as well as a number of shorter spans at each
end of the structure. The depth of the segments used in the central spans
are 8.25 feet and 14.75 feet at midspan and at the piers respectively. This
results in depth-to-span ratios of 1 to 17.5 at the pier and 1 to 31.6 at
midspan. This structure was constructed using precast segments erected
with a specially constructed gantry.
The typical cross section of the segmental alternate design of the
Napa River Bridge is shown in Fig. 4.48. This bridge has a main span of
250 feet and depth-to-span ratios of 1 to 32.26 at midspan and 1 to 20.83
at the piers adjacent to the longer spans. The bridge also has 7 spans of
150 feet each which have a constant depth of 7.75 feet. (Depth-to-span
ratio of 1 to 19.4.) Lightweight concrete was specified for the Napa
River Bridge.

3’-2” 14-9” 14-9”



Fig. 4.47 Ol&on Bridge between the European Continent and the island of OMon in France.

k 4

LVARIES (7’-9” MIN. TO 12-O” MAX.)

Fig. 4.49 Typical cross section of the Napa River Bridge, California.


Fig. 4.49 Typical cross section of the Saint Andre de Cubzac Bridge in France.

The viaduct at Saint Andre de Cubzac is an interesting example of a

bridge superstructure which has transverse floor beams in the deck. The
cross sectional dimensions of the segments are shown in Fig. 4.49 from
which it will be observed that the segments are quite wide (54.5 feet) and
have only two webs. The provision of the transverse floor beams results
in a relatively thin deck which spans longitudinally rather than trans-
versely. This structure has central spans of 312.5 feet and depth-to-span
ratios of 1 to 17.3 and 1 to 33.5 for the pier and midspan sections respec-
tively .
As mentioned above, bridge superstructures can be formed of two or
more tubular girders which are connected together by a deck slab as
shown in Fig. 4.50. This scheme can be used to facilitate providing
roadways of variable width. For superstructures of constant width, this
scheme permits the use of segments which are smaller, and hence easier
to cast, transport and erect, than would be the case if a single full-width
segment were used.
Because of the unique distribution of moments that exists in a
superstructure that is erected with the cantilever technique, the desira-
ble end span length of multi-span bridges is of the order of 65 to 70% of
the length of the interior spans. If shorter end spans are used, uplift can
be experienced at the abutment end of the end span due to live load on
the first interior span. Hence, special (and usually costly) support details
must be provided at the ends of short end spans so that the ends can
move neither up nor down. End spans longer than 70% of the interior
spans can of course be used with some sacrifice in economy.


r 1

Fig. 4.50 Bridge superstructure incorporating two tubular girders.


4.6 Proportioning the Segment

It is interesting to compare typical cross-sections of conventional cast-

in-place box girder bridges such as have been used so effectively in the
United States to those of typical segmental bridges of similar overall
width and span length. Such a comparison is made in Fig. 4.51 for a
bridge with a span of 150 feet. It will be observed that the depth-to-span
ratio of the conventional cast-in-place superstructure is of the order of 1
in 25 while that for the segmental bridge is 1 in 18.8. There is a simple
explanation for this difference. Firstly, the segmental bridge is erected in
cantilever and hence is subjected to a greater dead load negative mo-
ment than is the conventional cast-in-place superstructure. In addition,
the bottom flange of the segmental bridge, which is the compression
flange in areas of negative moment, has less width than is provided in
the conventional structure. Hence, more depth is required to compen-
sate for these factors.
In establishing the dimensions of the cross-section of a segment, one
must consider factors in addition to the longitudinal flexural stresses.
The flexural stresses are of course a major consideration but considera-




Fig. 4.51 Comparison of typical cross sections for cast-in-place box-girder and segme:tal-
girder bridges.

tions other than the longitudinal flexural stresses may control the dimen-
sions of a cross-section. If one examines the function of each of the
components of the cross-section, the following observations will be
made :
Flanges With respect to the longitudinal bending moments, the width
and thickness of the flanges, as well as the distance between them, have
direct bearing on the magnitude of the flexural stresses. In other words,
the dimensions of the flanges and the distance between them must be
great enough to confine the flexural stresses within the allowable limits.
This accounts for the fact that the bottom slab thickness frequently must
be greater in the vicinity of the piers than at midspan. The large negative
dead load moment which exists at the pier makes this factor especially
important when cantilever erection procedures are used. The bottom
flange must also be of sufftcient thickness to resist the transverse dead
and live load moments to which it is subjected. The top flange must act
as the deck and hence be of adequate thickness to resist the transverse
moments and shears due to the dead and live loads. An important prac-
tical consideration is that the flanges must be sufftciently thick to ac-
commodate the reinforcing steel and prestressing tendons that must be
embedded in them. A minimum thickness for the bottom slab of 5.5
inches is frequently specified while for the top slab a minimum thickness
of 6 inches is often used.
Web Layout The number and location of the webs has significant influ-
ence on the transverse bending moments in the deck as well as in the
bottom slab. Optimum deck design is achieved when there is a good
balance between the effects of the minima moment which occurs in the
deck at the location of the web and the maxima moment which occurs in
the deck between webs. This balance is not only affected by the number
and spacing of the webs but also by the thickness of the deck at midspan

L-9 aL

4 Fig. 4.52 Top slab proportions.

1 2 8 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

and at the support. The designer has the option of determining the
amount of slab overhang he wishes to provide in a particular design.
This is illustrated in Fig. 4.52 in which it will be seen that coefficient
“ a“, which is a measure of the slab overhang, is a major factor in
achieving an efficient design for the deck. The designer also must estab-
lish the thickness of the deck at the supports as well as at intermediate
points and in the deck overhang. These decisions also affect the efft-
ciency of the deck design.
Exterior webs are often inclined as a means of attempting to enhance
the appearance of a segmental bridge. This is illustrated in Fig. 4.53. It
will be recognized that inclining the web reduces the transverse span of
the bottom slab. This in itself is not of major importance. The fact that
the bottom slab is reduced in width by using inclined webs is an impor-
tant consideration in areas of negative moment. In areas of negative
moment the bottom slab is the compression flange. It is frequently
necessary to increase the bottom slab thickness in areas of negative
moment as a means of confining the longitudinal flexural stresses to an
acceptable level. A narrow bottom flange will require a greater increase


‘(aL ) bL aL

. L

Fig. 4.53 Cross sections with sloped and vertical webs.


k 8’-0” --3



Fig. 4.54 Web stiffeners and internal anchorage blocks.

in thickness to control the compressive flexural stresses than would be

required for a wider bottom flange.
Web Thickness A primary function of the webs is to resist the longitud-
inal shear stresses. Because these stresses are the greatest near the
piers, webs of superstructures having constant depth are frequently
made thicker in the vicinity of the piers than they are at midspan. Thick-
ening of the webs near the piers to accommodate shear stresses is not
normally required in superstructures of variable depth due to the reduc-
tion of the shear force acting upon the webs that results from the vertical
component of the compressive force in the inclined bottom flange. This
is known as the R&al effect. (See Section 5.2.) The reduction in the
shear force acting upon the concrete section as a result of the curvature
of the longitudinal tendons should be taken into account in bridges of
constant as well as variable depth.
For segments having cross sections which are symmetrical about a
vertical axis and which have only two webs, the shear force resulting
from the dead and live loads, as well as from the longitudinal prestress-
ing, is equally divided between the two webs. For segment cross-
sections which are not symmetrical about a vertical axis or which have
more than two webs, the distribution of the shear forces to the webs
1 3 0 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

may not be equal and must be determined using special methods. This
was explained in detail in Section 3.5. The webs also must resist flexural
stresses which are induced by transverse dead and live loads. In estab-
lishing the web thickness that is to be used in a particular design, in
addition to consideration of stresses, one must also consider the space
required to accommodate the embedded items as well as to provide for
the necessary minimum concrete thicknesses over and between the em-
bedded items. The provision of sufficient space for placing the concrete
during construction of the segments is very important. If the end an-
chorages of the longitudinal prestressing tendons are to be terminated in
the web at the joints between the segments, as has been done exten-
sively in the past, the web must also perform the function of an end-
block. The web thickness may have to be made greater than would be
required by other design considerations when the end anchorages are
placed in the webs.
Web Stiffeners A relatively recent innovation in precast segmental
bridge construction has been the provision of web stiffeners inside the
segment as shown in Fig. 4.54. The stiffeners are provided in order to
achieve a convenient means of anchoring temporary prestressing ten-
dons that are used in some erection techniques as well as to accommo-
date permanent prestressing anchorages. They can also be used to
strengthen precast segments with respect to stresses which are induced
during transportation and erection.
Fillets The fillets which are provided between the webs and flanges
serve an important function in relieving stress concentrations which
could result from transverse flexure in these locations. They also serve
the important function of providing necessary space for longitudinal pre-
stressing tendons together with longitudinal and transverse reinforcing.
An illustration of the concentration of tendons in the joint between a
web and top flange is given in Fig. 4.25.

Fig. 4.55 Forces on a joint between precast segments during erection.

Fig. 4.56 Multiple shear keys, (Courtesy of Sob&6 Technique pour L’Utilization de la
Pr&ontrainte, Paris, France).

Shear Keys Shear keys are provided in the webs and flanges of precast
segments. They serve two functions. The first is to align the segments
when they are erected. The second is to transmit the shear force be-
tween segments during construction when the epoxy bonding compound
applied to the segment joint is still plastic and behaves like a lubricant.
The web shear keys alone must be relied upon to transfer the shear force
across the joint. The plastic bonding compound precludes the existence
of friction between the segments. The forces acting on a joint during
erection are as shown in Fig. 4.55. Although in the past the epoxy bond-
ing compound provided in the joints between precast segments has been
relied upon to transmit shear stresses in the completed structure, the use
of multiple shear keys as shown in Fig. 4.56 makes this unnecessary.
The web shear keys should be designed using the stresses permitted in
Chapter 17 of the Building Code Requirements for Reinforced Concrete
(Ref. 7).
Anchorage Blocks In precast segmental construction the longitudinal
prestressing tendons are frequently anchored inside of the segments at
1 3 2 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

internal anchorage blocks as shown in Fig. 4.54 rather than in the webs
at the joint between segments as has been traditional in cast-in-place
construction. Two advantages evolve from this practice. Firstly, the
webs do not have to function as end blocks and resist the high tensile
stresses that occur in the concrete which is immediately under the end
anchorages. Secondly, this procedure permits the construction crews
that are installing and stressing the permanent prestressing tendons to
work independently of the crews that are erecting the segments and
installing the temporary prestressing. Internal anchorage blocks are
frequently, but not always, accompanied by web stiffeners which were
described above.
Segment Length In designing and detailing a bridge that is to be con-
structed segmentally, one needs to know the length of the segments in
order to completely analyze the structure and check the stresses in the
joints between the various segments. One cannot be certain his design is
complete unless this is done. On the other hand, the designer generally
prefers to leave decisions such as this to the contractor with the purpose
of not being unnecessarily restrictive and obtaining the lowest cost for
the owner. One solution to this problem is to select a segment length
that seems reasonable under the circumstances and prepare the com-
plete design on this basis. An appropriate clause in the specifications
can be provided to allow the contractor to use other than the detailed
segment length and prestressing details providing he submits proof of the
adequacy of any changes he wishes to make.
With these thoughts in mind, one should consider the factors which
dictate segment length. A primary consideration influencing the size of
the segment is the unbalanced moment the substructure, substructure-
superstructure connection, or substructure foundation, can withstand.
In cast-in-place construction, the design of the traveling forms is af-
fected by the weight of the segments they support. One large company
that has constructed many cast-in-place segmental bridges has devised a
standard adjustable traveler that will accommodate a segment having a
maximum length of 16 feet. This then seems like a reasonable upper
limit for cast-in-place segments. Precast segments which must be hauled
over the public highways, should not exceed lengths and weights that
would require special fees, transportation costs or special permits. A
length of 8 feet would normally be able to be hauled without exceeding
legal load widths but segment weights that can be accommodated must
be investigated for each locality. Precast segments which are to be job-
site cast and not hauled over public roads or highways are limited only
by the practical considerations of fabrication, handling and erection. For
segments which are to be cast-in-place on falsework, the practical limita-

tions of available falsework, construction sequence and span lengths will

dictate the lengths of segments to be used. Needless to say, not all con-
tractors who bid any one job will agree as to the optimum segment
From the above it will be apparent that many basic considerations
other than stresses resulting from longitudinal bending should be taken
into account in establishing the cross-sectional dimensions of the seg-
ment. Some of these considerations are interrelated. If, for example, it is
found that the thickness of the webs or flanges of a section must be
increased in order to accommodate the necessary embedded items or to
properly control transverse bending stresses, the dead load of the struc-
ture will be increased as will be the stresses resulting from longitudinal
dead load bending moments and shears. It should also be apparent that
the method of erection that is to be used can have a very significant
influence on proportioning segments for longitudinal flexural stresses.
Hence, it should be apparent that proportioning for longitudinal bending
involves the commonly used procedure of adopting a trial section which

. SPAN = L



Fig. 4.57 Comparison of the elastic curves for no hinge within a span to those for one at ’
midspan and one at the quarter point.
1 3 4 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

has dimensions that are assumed to be adequate for the conditions at

hand and reviewing the adequacy of the assumed section. If in-
adequacies are found, the dimensions of the section are adjusted as
necessary and the process is repeated until a complete and satisfactory
solution is developed.

4.7 Intermediate Hinges

Variation in length due to the effects of temperature, creep and shrink-

age must be accommodated in bridge design. This factor becomes in-







Fig. 4.58 Procedure in providing an intermediate hinge within a span at other than midspk.

creasingly important with the length of the bridge. This is most often
done by providing joints within one or more spans which include provi-
sion for translation and rotation. The joints are referred to as hinges.
The provision of a hinge within a span results in a discontinuity of the
structure. Hence, abrupt angles between the tangents to the structure on
each side of the hinge may develop due to applied loads or the effects of
creep. Elastic curves for three hinge conditions in an interior span,
which is a part of a multi-span continuous beam, are shown in Fig. 4.57.
The curves are for the condition of a uniformly distributed live load with
a hinge located at mid-span, at quarter point and without a hinge. Note
that the angle between the tangents to the elastic curve on each side of
the hinge is nearly three times as great when the hinge is at midspan than
when it is at the quarter point. Hinges should be avoided where possible.
When required, it is better to place them at the quarter point than at
! Hinges can conveniently be provided between midspan and the sup-
port at any desired location in segmental construction. This is done by
1 providing temporary prestressing and blocking across the hinge until
such time as the abutting cantilevers are joined at midspan. The hinge is
rendered free to translate and rotate by the removal of the temporary
prestressing tendons and blocking. This procedure is illustrated in Fig.
It should be recognized that releasing a hinge which was fixed tem-
porarily during construction results in a redistribution of the moment
that existed at the joint while it was restrained. The effect of releasing
the hinge on the stability of the structure, with the configuration it has at
the time the hinge is released, must be investigated. At the time the
hinge is released, the portion of the structure from the hinge back to-
ward the direction from which the construction proceeded is continuous
while that portion from the hinge toward the direction the construction is
proceeding may be simply supported and unable to support the reaction
from the continuous structure without auxiliary supports. In addition to
the moment which results from the releasing of the hinge, moments due
to the removal of temporary hinge prestressing tendons must be taken
j into account when the hinge is released.

4.8 Support Details

Tall slender piers are frequently relatively flexible in comparison to the

superstructure and hence the moments induced in the piers due to vertical
loads as well as due to superstructure length and rotation variations are

relatively small. The result is that the taller piers can frequently be con-
structed with a moment-resisting (fixed) connection to the superstructure.
On the other hand, short stiff piers frequently must be provided with
bearing devices in the joint between the top of the pier and the superstruc-
ture. Properly selected bearing devices permit the designer to control the
moments and shears that are induced in the pier due to vertical loads,
superstructure length variations and horizontal forces. This, of course, is
true in bridges of all types and is not unique to segmental bridges.
In segmental bridges, the design of the superstructure segments over
the piers must be such as to be able to transfer the required forces and
moments between superstructure and substructure.
Consider the pier proportioned with a width that is less than the width
of the superstructure soffit as shown in Fig. 4.59. It is apparent that the



Fig. 4.59 Pier with a width less than the width of the superstructure



Fig. 4.60 Pier with a width equal to the width of superstructure





Fig. 4.61 Pier cross section.



. i-

\ ~_ _ ---
’ 7

Fig. 4.62 Pier cross section

diaphragm provided in the pier segment must be suffkiently strong to

transfer the vertical loads from the superstructure webs to the pier. If
the connection is designed to fix the top of the pier and the superstruc-
ture with respect to rotations, the diaphragm must also be capable of


Fig. 4.63 Pier cross section.

withstanding the torsional moments which occur between the
superstructure webs and the pier shaft. The torsional stresses in the
diaphragms could be avoided by using a pier that is as wide as the
superstructure sofftt as shown in Fig. 4.60.
The cross-sections of three pier shafts are shown in Fig. 4.61, 4.62
and 4.63. The cross section of Fig. 4.61 consists of two intersecting
wall-like elements. The short elements which project along the center
line of the bridge may be considered desirable from an architectural view-
point but such projections can seriously restrict the moment capacity
of a section and can create special problems in effecting a moment trans-
fer between the superstructure and substructure. The cross section
shown in Fig. 4.62, although it may be less desirable from an appearance
standpoint, is far more efftcient structurally than that of Fig. 4.61. This
is particularly true when moment transfer between pier and superstruc-
ture is to be made. A pier cross section such as is shown in Fig. 4.63 is
offered as one of many possible compromise solutions for obtaining a
structurally efficient section that still provides protrusions and hence
shadows to enhance the appearance of the finished structure. Piers
which are of variable dimension, as shown in Fig. 4.64, are sometimes
used in an effort to enhance the appearance of a structure. The variable

Fig. 4.64 Pier with variable dimension. (Courtesy California Department of





Fig. 4.65 Wide pier for moment transferring ability.



Fig. 4.66 Wide pier for moment transferring ability.


dimensions are not normally desirable from the standpoints of cost and
structural efficiency.
Supports which have a reasonable width in the direction of the span at
the junction with the superstructure, say of the order of eight to ten feet,
as shown in Fig. 4.65, facilitate superstructure erection when the can-
tilever erection method is to be used. When the width is adequate, the
unbalanced moment which occurs at the support with the cantilever
erection technique can be transfered with temporary prestressing ten-
dons together with temporary bearing pads. Provision of adequate space
and means of access for inserting hydraulic jacks, as shown in Fig. 4.66,
permits removal of the temporary bearings and insertion of the perma-
nent bearings upon the completion of the cantilever. Provision of space
and access for the hydraulic jacks also permits adjustment of the posi-
tion of the completed cantilever before the continuity joint between ad-
jacent cantilevers are cast. After final adjustment of the cantilever and
insertion of the permanent bearings, the connection can be left free to
rotate, and hence act as a hinge, or can be fixed against rotation by
grouting the joint and inserting permanent prestressing tendons across

4.9 Construction Details

When using the cantilever erection technique together with tendons com-
posed of high strength wires or strands, tendon layouts similar to those
shown in Fig. 4.67 are frequently employed. The tendons are generally
referred to as “cantilever” tendons and “continuity” tendons.
The cantilever tendons are those which are located in the areas of
negative moment and are normally placed and stressed during the erec-
tion of precast segments or casting of cast-in-place segments. They are
generally on a trajectory that is symmetrical about the centerline of the
pier. Cantilever tendons may be extended downward from the upper
surface of the structure as shown in Fig. 4.67 or may terminate near the
upper surface of the structure. When extended downward into the webs
of the segments, the cantilever tendons exert a vertical component of
prestress which reduces the shear force imposed upon the concrete sec-
tion. When the tendons (either cantilever or continuity) are extended
into the webs, the French Code (Ref. 50) requires the shear design be
based on the net section.
The continuity tendons are those that provide prestressing in the
areas of positive moment. They are placed and stressed after the closure
joint has been placed. The continuity tendon may or may not be ex- ’
1 4 2 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S


jI N T E R I O R S P A N

Fig. 4.67 Typical layout of longitudinal wire or strand tendons with the CantileVer
erection technique.


Fig. 4.66 Tendon layout for tendons which do not extend into the webs.
S E G M E N T A L B O X - G I R D E R B R I D G E S 11 4 3

Fig. 4.69 Recommended bridge rail detail.

Fig. 4.70 Bridge rail detail that is not recommended.

1 4 4 (H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

tended up into the webs of the superstructure as shown in Fig. 4.67.

A tendon layout in which the continuity and cantilever tendons are
not extended into the webs is shown in Fig. 4.68. In the layout in Fig.
4.68 the tendons are shown to be anchored at anchorage blocks inside of
the segments. The cantilever tendons could be anchored in the ends of the
segments (at the joint between segments) if the internal anchorage blocks
are not provided.
When bar tendons are employed, the prestressing force is varied by
terminating bars at various locations along the span in much the same
manner as is done in reinforced concrete. Curvature of bar tendons is
kept to a minimum due to the fact it is much easier to place bar tendons
on straight trajectories.
The webs of segmental bridges are sometimes prestressed as a means
of reinforcing for shear. This may be accomplished to various degrees
by extending wire or strand tendons into the webs as shown in Fig. 4.67.
It can also be accomplished by placing tendons, either vertically or
diagonally, in the webs. Some engineers feel that the webs of all long-
span concrete bridges should be prestressed as a means of controlling
shear-induced web tensile stresses. It must be recognized that many exist-
ing concrete bridges having webs reinforced with non-prestressed web
reinforcing are giving very satisfactory service.
Internal anchorage blocks, together with “web stiffeners” as shown in
Fig. 4.54 are sometimes employed in precast segmental construction as a
means of avoiding terminating the tendons at the joints between seg-
ments and providing a means for attaching temporary prestressing ten-
dons. These details have been found to be useful in reducing the time
required to erect precast segments. Basically it permits the workmen
who are erecting the segments to work almost independently of those
installing and stressing the cantilever tendons.
The detail used for the bridge barrier rail deserves mentioning. The
recommended detail is shown in Fig. 4.69. It is characterized by the fact
the cast-in-place railing extends below the edge of the deck. In this
manner any difference in grade between the ends of cantilevers which
may exist at the cast-in-place closure joints, will be hidden. The detail
shown in Fig. 4.70 is obviously less costly because the forms are more
easily erected and stripped. This is because they do not extend below
the top of the deck. The latter is less desirable from an appearance
standpoint for the reason previously mentioned.
5 Additional

5.1 - Introduction

Included in this Chapter are brief discussions of several topics which may
be completely understood by the reader and hence need not be given
furtherconsideration. On the other hand, experience has shown that not all
bridge designers are familiar with these subjects. It is for this reason that
this chapter has been included.

5.2 - Design for Shear

Contemporary shear design provisions of the AC1 31% Building Code

Requirements for Reinforced Concrete (Ref. 51) for prestressed concrete
flexural members involves mathematical relationships that are more com-
plex than the relationships contained in the AASHTO Specification (Ref.
6). The AASHTO Specification permits the use of the ACI criteria. The
current provisions of AC1 318 may more accurately reflect the true shear
strength of prestressed concrete than was predicted by previous design ’

criteria. Many engineers, however, question the necessity for the more
complex relationships and point out that the new relationships do not show
obvious economy with respect to shear design nor were the previously
used relationships obviously inadequate.
The fact that some engineers believe the AC1 criteria are unconservative
when applied to large concrete beams is worthy of note. This group of
engineers is of the opinion that when the shear capacity of a concrete beam
is exceeded by the design load, the entire shear force should be carried by
reinforcing steel. They believe the combined effects of dowel action,
aggregate interlock and friction in the compression flange can be significant
in the shear strength of small beams but not significant in large beams (Ref.
52). They point out the fact that shear strength relationships contained in
AC1 318-71 have been confirmed by results obtained from tests of small
The contemporary shear design criteria of AC1 318-71 lends itself to
computer analysis. A mini-computer or programmable calculator of mod-
erate size will suffice. Once the computer has been properly programmed,
the fact that the computations are somewhat complex is unimportant. One
cannot, however, properly program a computer unless he completely
understands the meaning of the terms which are included in the relation-
ships involved.
There is considerable confusion in the meaning of the terms in Eq. 1 l-l 1
of AC1 318-71. This equation, as revised in 1973, is:

vc, = 0.6 c + Vd + i):trJ (5.1)

The official AC1 definitions of the terms in the second part of the equation,
are as follows:
Vd = Shear force at section due to dead load.
Vi = Shear force at the section considered due to externally applied
design loads occurring simultaneously with Mm=.
MU = Bending moment causing flexural cracking at the section consi-
dered due to superimposed loads.
M max = Maximum bending moment at the section considered due to
externally applied design loads.
bw = Web width, or diameter of circular section, in.
d = Distance from the extreme compression fiber to centroid of
reinforcement. in.

It is the opinion of the authors that the definition of Vd should be the shear
force at section due to service dead load because it is not intended that this

load be factored. The term “externally applied design loads” which is used
in the definition of Vi and Mmax should be defined as the design loads less
the service (unfactored) dead load. It should be recognized that design
loads are factored service loads.
It should be remembered that Vci is intended to be a measure of flexurally
initiated shear cracks while vcW is intended to be a measure of cracks
resulting from excessive principal tensile stress. The relationship for vcw
(Equation 11-12 of AC1 315-71) is as follows:

vcu = 3.5 fl + 0.3 fp + JL (5.2)


in which VP is the vertical component of the effective prestresdg force at

the section being considered and fpc is defined as follows:
fpc = compressive stress in the concrete, after all prestress losses
have occurred, at the centroid of the cross section resisting the
applied load or at the junction of the web and flange when the
centroid lies in the flange, psi. (In a composite member, fP will
be the resultant compressive stress at the centroid of the com-
posite section, or at the junction of the web and flange when the
centroid lies within the flange, due to both prestress and to the
bending moments resisted by the precast member acting alone.)
As an alternative to Eq. 5.2, one may compute the value of vcW as being the
shear stress which results in a principal tensile stress of 4 vr (3 \/fc’
should be used in lieu of, 4 fl for all-lightweight concrete while 3.4 v’F
should be used for sand-lightweight concrete). For a member stressed
longitudinally only, this can be expressed as follows:

v,, = J4vfZ(4UP +fW)+$ (5.3)

In the case of a member having a vertical prestress f& in addition to a

longitudinal prestress of fpc, the shear stress vcw which results in a principal
tensile stress of 4 fl can be determined from:

v,, = 44 VP + f’p, (4G + f,) + 2 (5.4)

Although Eqs. 5.3 and 5.4 would have been considered much too difftcult
for every day use in the day of the slide rule, they are easily solved today on
inexpensive electronic calculators.
It has been pointed out that prestressing tendons which are inclined to L
the gravity axis of a beam have a vertical component offorce which usual/y

is acting opposite in direction to the flexural shear force. Hence, the

vertical component of the prestressing force should logically be taken into
account when computing the actual stress on the concrete. Equation 5.2,
which is an “allowable stress,” includes this effect as does Equation 5.4. It
should be observed that the relationship for Vci (Eq. 5.1) does not include
the effect of the vertical component of prestressing. The explanation for
this given by AC1 Committee 426 is that the effect of VP is included in the
flexural cracking moment, Mcr (which is included in Eq. 5.1) and hence VP
does not have to appear in Eq. 5.1 (Ref. 53). In continuous bridge struc-
tures VP can be quite large in areas of negative moment and it seems logical
that its effect be incorporated in the analysis. Research is needed on large
continuous prestressed concrete beams with the objective of obtaining a
better understanding of the shearstrength of such members in the areas
subject to negative moment.
Beams of variable depth and normal configuration have less shear force
on their webs in areas where the compression flange is inclined to the
gravity axis than would be revealed from a usual analysis of the flexural
shear forces. The principle is illustrated in Fig. 5.1 in which a freebody


Fig. 5.1 Freebody of a portion of a bridge superstructure having variable depth. ‘(R&al

diagram of a portion of a variable depth continuous beam is shown. The

portion of the beam shown is near the support where both the shear force
and the negative moment are large. If the angle of inclination of the bottom
slab with respect to the gravity axis of the member is taken as (Y and the
force in the compression flange is designated as C, there is a vertical
component of the force C. The vertical component is equal to C since. The
vertical component of the force C acts in a direction that reduces the shear
force acting on the webs of the member. When applying this effect in an
analysis under design loads, the force C should be determined based upon
the design moment that is concomitant with the design shear force being
Secondary moments and shear forces frequently exist in continous pre-
stressed concrete members. It is recognized that the effect of these should
be included in strength calculations (Ref. 54). A load factor of unity is
generally accepted as being appropriate when including these effects in
strength calculations.
In view of the above, it would seem the provisions for shear design of
prestressed concrete members, following the principles of AC1 318, be
done by including the effects of variable depth and secondary reactions
(with load factors of unity) in the computation of the design shear force at
each section considered. Hence, with the AASHTO load factors, this
relationship becomes:

v,, = 1.44VD + 2.41 (V, *,) + V” + vs (5.5)

in which
V,D = Total dead load shear
VL+I = Shear due to live and impact loads
VV = Effect of variable depth
VS = Shear due to secondary effects of prestessing

, The values of Vci and vcw would then be computed using equations 5.1 and
In using the lane load design criteria of the AASHTO Specification, it
will be remembered that the magnitude of the concentrated loads used in
determining maximum values of moment and shear differ. Furthermore, in
continuous spans, two concentrated loads are used in determining the
maximum negative moment. Only one load is used in determining
maximum positive moments (even in continuous structures) and maximum
shears. Because the shear Vi is the shear at the section under consideration
when the loads are ofthe magnitude andpositioned to produce the greatest
value of Mmor, the correct value of Vi is less than would be obtained if the ’
1 5 0 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

largest live load shear force which occurs at the section were used in
determining L’;. It should also be recognized that computing Vi using the
largest live load shear force which occurs at a section, rather than the shear
force which is conco’mitant with the loading which produces Mm+ is
unconservative because it will result in a greater value of Vci.

5.3 Horizontally Curved Bridges

Bridges which have closed cross sections, such as the box-girder bridges
described in Chapter 3 and the segmental bridges described in Chapter 4
are efficient in resisting torsional as well as flexural moments. For this
reason bridges of these types are especially desirable in structures which
must have a horizontally curved alignment. The horizontal curvature re-
sults in torsional moment from both dead and live load. Intermediate
diaphragms are generally provided in box-girder bridges that have signific-
ant horizontal curvature. Diaphragms are normally required at each sup
port in order to transfer shears and moments from the superstructure to the
k method for the rigorous analysis of a horizontally curved structure,
which requires the use of a computer, is described in the literature (Ref.
55). This rigorous method has been used to confirm the adequacy of an
approximate method for usual design applications. The approximate
method, which has been described by Witecki (Ref. 56) is simple to apply
and does not require a computer.
The approximate method described by Witecki is based upon the rela-


in which
M = total external bending moment produced by dead load,
superimposed dead load, live load and any other anticipated
load-all computed as for a straight structure.
R = radius of curvature.
M/R = torque per ft. produced by bending moment.
T = qe = torque per ft. produced by eccentrically applied load.
S = length on which the torsional moment is being computed.
The method is used to determine the torsional moments as follows:
1. Determine the vertical dead, superimposed dead and live load.
2. Determine the uniformly applied torque moment which must be ac-
counted for in the design. This torque moment should include the

effects of dead and live load eccentricity as well as centrifugal force

and wind.
3. Assuming the bridge is on tangent (without horizontal curvature)
compute the longitudinal bending moments for the applied loads.
4. The summation of the bending moments at each section should then
be divided by the radius of curvature and the quotient added to the
torque determined in step 2. The result so obtained is the term z + T
for each section.
5. The spacing of the supports along the structure which are capable of
restraining the torsional moment determines the torque spans. If the
piers are incapable of restraining the torsional moment, all of the
torsion must be resisted at the abutments and the length of the bridge
becomes the torque span. The results of step 4 are applied as a
distributed load on the torque spans, which are assumed to have
simple supports, and the beam reactions and distribution of shear
forces are determined. The shear forces so determined are equal to
the torsional moment at the various sections and the reactions are the
concentrated torsional moments which the supports must resist.
Where two torque spans frame into a single support, the support must
restrain the torsional moment that is equal to the algebraic sum of the
reactions from each span.
6 . The longitudinal shears and bending moments are determined as ifthe
bridge were on tangent. It has been found that the horizontal curva-
ture has little influence on them under normalconditions (Ref. 57,58).
An easily followed numerical example for the approximate method can
be found in Reference 56.

5.4 Strength Analysis

Continuous structures, particularly those which have short spans and rela-
tively large moving live loads, can experience moment reversals in un-
loaded spans which are adjacent to loaded spans. This phenomenon is well
known and is not unique to prestressed concrete structures.
A related strength consideration in the design of prestressed concrete
continuous structures which is not as well known is illustrated in Figs. 5.2
and 5.3. Moment capacity envelopes are shown in the two figures together
with a moment diagram that is associated with a particular combination of
design loads. In Fig. 5.2 the moment diagram is for the condition of
‘secondary moment due to prestressing plus design dead and design live
loads on all the spans. For the moment diagram shown in Fig. 5.3, the
loading consists of secondary moment due to prestressing plus service *

dead load on all the spans together with design live load on alternate spans.
It will be seen that with design dead and design live load on all the spans,
the simple span moment capacity for each span is 46,300 kip-Et. and the ,
negative moment capacity of -35,400 kip-ft. at each support can be fully
utilized. With service dead load on all spans together with design live load
on alternate spans the simple span moment capacity for the loaded span is
34,200 kip-ft. In the latter case the negative moment capacity at the ends of
the loaded spans is restricted to -23,300 kip-ft. due to negative moment
capacity of -5,200 kip-ft. at midspan of the unloaded spans. It will be
recognized that when the moment of -5,200 kip-ft. is attained at midspan of
the unloaded spans, plastic hinges will form. Further increase in load will
increase the positive moment at midspan of the loaded span until the value
of 10,900 kip-ft. is obtained at which time an additional plastic hinge would
form and the structure will no longer be stable.,
This type of strength analysis should be made for all continuous pre-
stressed concrete bridges. It is especially imperative that this type of
analysis be made for segmental bridges which are erected in cantilever.
This is because there is a tendency in this type of structure to have little, if
any, negative moment capacity at midspan. It is frequently found to be
necessary to place steel that is not required at service load in the upper


// I \\ I /I I \\ I! -[/// I
! +10,900!

Fig. 5.2 Ultimate moment capacity diagram for a continuous prestressed concrete struc!ure
under design dead and live loads on all spans. (After J. Muller Ref. 67.)


+lo-goo L D.‘L. + FACiORED L.L.I

Fig. 5.3 Ultimate moment capacity diagram for a continuous prestressed concrete structure
under service dead load on all spans and design live load on alternate spans. (After
J. Muller Ref. 67.)

Fig. 5.4 Elastomeric bearing pad used to explain European design practice.
154 ) H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

portion of the section near midspans in order to provide the required

ultimate moment capacity.
In girder bridges formed of precast prestressed concrete beams in com-
bination with acast-in-place deck which contains continuity steel, positive
moments can sometimes develop near the supports as a result of tempera-
ture, creep and shrinkage as well as due to live loads in remote spans.
Provision of non-prestressed reinforcement is recommended in these areas
as a means of controlling the effect of these moments.

5.5 Elastomeric Bearing Pads

It is interesting to compare the criteria contained in the AASHTO
Specification for the design of elastomeric bearing pads to that which is
used in Europe (Ref. 59). Such a comparison is made in Table 5.1. The
obviously important differences are in the allowable compressive stress
and the allowable translation.
The European design criteria can be explained by considering a pad
having the dimensions shown in Fig. 5.4. The following notation is used:

G = Shear modulus of the elastomer

n = Numbers of layers
a = Dimension parallel to the longitudinal axis of the bridge
b = Dimension perpendicular to the longitudinal axis of the bridge
t = Thickness of one layer of elastomer (between bonded steel plates)
T = Total thickness of the bearing pad T = nt
A = Area of the bearing pad
S = Shape factor = 2t (a + b)
TP = Shear stress due to vertical load
7 = Total shear stress
7h = Shear stress due to horizontal load = TS + rd
7s = Shear stress due to static horizontal load
‘Td = Shear stress due to dynamic horizontal load
7a = Shear stress due to rotation
cl = Normal stress
6s = Horizontal deflection due to static Joads
$d = Horizontal deflection due to dynamic loads
HS = Static horizontal load
Hd = Dynamic horizontal load
The basic consideration in the design of elastomeric &afing pads is
the shear stress which exists in them due to the action of vertical load,

Allowable compressive stress
D. L. 500 psi none
T. L. 800 psi 2000 psi
Allowable shear stress
Total none 570 psi
Vertical load (approx.) none 340 psi
Horizontal displacement none 60 psi
Rotation none 170 psi
Translation .
Long term (temp., shrinkage, creep) none 0.5 T
Long term (temperature) See below
Short term (Dynamic loads) none 0.7 T
Rotation 1
For each layer cq 3 0 -a 2

For total a.~ 3n0 1; *

VerIica[ Deflection
T. L. 0.07 t none
Edge Clearance 3 in. 2 in.
The AASHTO Specification limits the total of the positive and the negative movements to
0.5 T or less.

horizontal load and rotation. it should be recognized that the application

of a vertical load on a solid block of elastomer results in bulging of the
sides of the block as shown in Fig. 5.5. The bulging is accompanied with
shear stresses. A commonly used means of controlling the shear stresses
(and reducing the bulging) is to provide pads composed of several thin

Fig. 5.5 Bulging of an elastomeric bearing pad due to vertical load



: A

Steel Plates

Fig. 5.6 Reinforced elast!meric bearing pad.

layers of elastomer bonded to thin steel plates separating the layers of

elastomer. This is illustrated in Fig. 5.6.
The total shear stress in the elastomer, 7, should be limited to 570 psi.
Under the vertical load, the shear stress is:

Tp = - 3P (5.71
2 abS

If TP is limited to a maximum value of 340 psi, the maximum value of the

vertical stress is 225 S psi where S is the shape factor defined above. The
normal stress is generally limited by the concrete bearing stress. Because
only nominal reinforcement is generally found in the concrete near the
bearings, this stress is limited to 2000 psi and a minimum edge distance of 2
inches is normally used.
Shear stresses due to static and dynamic horizontal loads must be
considered separately. Static loads include the effects of temperature
concrete shrinkage and concrete creep. Dynamic loads include the ef-
fects of live loads such as braking forces, traction, wind and earthquake.
The shear stress due to a static horizontal load is:

T, - G& - I-I, (5.8)

I- ab

and the maximum allowable value is 0.5G (60 psi) which limits the horizon-
tal deflection due to static loads to 0.5 T.
The shear modulus of the elastomer, is twice as great for loads of short
duration as it is for long-term loads. Hence the shear stress due to dynamic
loads is:
ZG 8, _ H,,
T,I = - - - (5.9)
T ab

which can be written

ad = Hd T (5.10)
2G ab
The combined shear stress due to horizontal loads is:

If the deformation of the bearing pad under horizontal loads is restricted

to an angle y whose tangent is 0.7,
tan y = $ (5.12)
and one can write

4 0.7 (5.13)

The shear stress due to rotation is equal to:

T,, =-
G ( a )‘tan
2 T
(Y (5.14)

and the maximum value is limited to 1.5 G. Because for small angles,
the tangent of the angle is equal to the angle, the rotation of the bearing
is limited to

for each layer of elastomer or

t 2 (5.16)
(~7 S 3n ( a 1
for n layers.
If the vertical stress o is less than 430 psi or if the total horizontal force
equals or exceeds 20% of the vertical force, the bearing will slide unless it is
bonded to the surface on which it bears.
Comparing the above design criteria to that shown from the
AASHTO Specification in Table 5.1, will reveal our European counter-
parts are using compressive stresses that are over two and one-half
times larger than those permitted by the AASHTO Specification. In ad-
dition, while the AASHTO Specification limits the sum of the positive
and negative displacements of the bearing pads to one-half of the height
of the bearing, the European criteria limits the maximum displacement
under the combined effects of static and dynamic loads to 0.7 of the
height of the pad.
It is interesting to note that elastomeric bearing pads were first used in
158 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S


G Pier


Fig. 5.7 Example of reactions at a pier which is provided with twin, spaced elastomeric

civil engineering works in Europe. The performance of the elastomeric

bearing pads designed with the European criteria has been good and
would indicate that it is adequately conservative.
The design criteria used in the U.S. results in bearing pads much
larger than are required by the European criteria. This is best illustrated
by an example. Consider the case shown in Fig. 5.7 in which plan and
elevation views of a bridge are shown at a bent. The detail shown con-
sists of two pads placed near the edges of the bent which is eight feet
wide. Spaced as shown, the dead load reactions are 2200 kips on each
pad and the live and impact loads result in an upward reaction of 905
kips on the side of the unloaded span and a downward reaction of 1335
kips on the side of the loaded span. The result is a maximum reaction of
3535 kips on one pad and a minimum reaction of 1295 kips on the other.
The AASHTO Specification would require pads having an area of 30.7
square feet while the European method would require pads having an
area of 12.3 square feet.
An increase in pad area requires a greater height for the pad if the pad
must accommodate a particular movement and if the force which causes
the pad deformation is to be unaffected by the increased area. From this
it will be seen that the low allowable stresses due to vertical loads result
in very considerably larger pads being required with the AASHTO

Specification as compared to that required with the European design

If properly detailed, the use of elastomeric bearing pads can allow the
bridge designer to control the forces a support must withstand from a
given translation. Increasing the height of the pad will reduce the hori-
zontal force transferred to the support. A decrease in the pad height
increases the force and the effective stiffness of the substructure. By
using spaced bearing pads as shown in Fig. 5.7, the designer can transfer
moment due to vertical loads from superstructure to substructure and
yet control the moment induced in the bent as a result of translations. In
effect, semi-fixed and semi-expansion bearings can be acheived by
proper proportioning. Bearings of other types, such as conventional
metal bearings and elastomeric “pot” bearings are not as versatile in
forming semi-fixed or semi-expansion bearings.
The spaced bearing detail shown in Fig. 5.8 was used on the Oleron
Bridge in France. The use of the spaced bearings resulted in a distribu-
tion of moments due to the design live load of four kips per linear foot as
shown in the left side of Fig. 5.9. If a single row of bearings rather than
two rows were used, the distribution of moments would have been as
shown in the right side of Fig. 5.9. From this it will be seen the provi-


J 4

Fig. 5.8 Superstructure-substructure detail at a typical interior pier of the 0l”ron Bridge in
1 6 0 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

sion of spaced bearings results in the span having a larger negative mo-
ment and smaller positive moment than it would have with a single
hinged support; the spaced bearings increase end moment restraint and
reduce deflections.
Consider the five-span bridge shown in Fig. 5.10. The structure
traverses a deep valley which results in the two interior bents being rela-
tively high while the outermost bents are relatively low. The prestressed
concrete superstructure will shorten with the passing of time due to ef-
fects of shrinkage and creep. In addition, its length will vary with temp-
erature changes. The changes in length require the provision of one or
more expansion joints in the superstructure. The number and location of
the expansion joints has significant influence on the moments which will
exist in the structure as a result of length changes as well as due to
lateral loads. Normally one prefers to use the minimum number of ex-
pansion joints possible because joints require maintenance during the life
of the structure and may be costly to provide.
One joint layout that could be feasible for the bridge of Fig. 5.10, would
consist of expansion joints at the abutments only. With this layout, if the
superstructure were fixed to the bents, the changes of length of the
superstructure would induce moments in the bents. The moments in the


4 260 260 266 -



Fig. 5.9 Distribution of moments in a typical interior span of the Okon Bridge with spaced
bearings and with a single row of bearings.

Fig. 5.10 Five-span bridge traversing a deep valley.

taller central bents are frequently not critical under such circumstances
whereas they frequently are critical for the shorter and stiffer outer bents.
Provision of expansion bearings between the top of the shorter bents and
the superstructure would eliminate the moments in these bents but would
require the higher bents to resist all longitudinal lateral loads.
Another and perhaps better method of solving the problem is to place
spaced elastomeric pads, as shown in Fig. 5.7, between the tops of the
shorter bents and the superstructure. By carefully selecting the size and
shape ofthe bearings, the designer can obtain almost any effective stiffness
of the combined bearing and bent and still transmit moment due to vertical
loads to the shorter bents. This is discussed further in the following section.

5.6 Substructure Considerations

The piers or bents of a bridge constitute an important part of the bridge

from functional, cost and appearance standpoints. Too frequently the
shapes of piers or bents are selected without each of these factors being
given the consideration it deserves. Each of these is treated briefly in this
The primary function of a pier is to transmit the vertical dead and live
loads ofthe superstructure to the ground through the foundation system. A
secondary function which piers sometimes perform consists of transmit-
ting lateral loads, both transverse and parallel from the longitudinal axis of
the bridge, to the ground. Another secondary function piers sometimes
perform is to resist moments induced in them either as a result of vertical
dead and live loads acting upon the superstructure or due to dimensional
changes (temperature, creep and shrinkage) the superstructure may un-
dergo. These latter functions are secondary in nature but not necessarily in
magnitude. Many times the bridge designer must take special precautions,
such as using elastomeric bearing pads as was described in Section 5.5, to
eliminate or reduce what could otherwise be excessive stresses.
Piers of various shapes and cross sections have been successfully used.
This includes simple solid round or rectangular columns as shown in Fig.
5.11, simple hollow prismatic shafts as shown in Fig. 5.12, twin walls which
may or may not be vertical as shown in Fig. 5.13 and flared shapes such as
is shown in Fig. 5.14, to show only a few examples. Each type has certain
characteristics and advantages which can be described as follows:

Solid piers. Solid round and rectangular columns have been extensively
used in bridge construction and most particularly in bridges spanning
roads, highways and railroads. Their use has generally (but not always)
been confined to structures of moderate span and height. Primary advan-
tages of these shapes include relatively low cost, unless they are large in
cross section, and acceptable appearance under most conditions.

Fig. 5.11 Solid piers which are round in cross section. (Courtesy California Department of

Fig. 5.12
Prismatic piers which are hollow.

Hollow prismatic piers. Hollow prismatic piers are structurally effkient,

are relatively low in cost and generally find application on structures of
moderate to long spans. They can be used to best advantage for piers that
are of medium to great height. Piers of this type are generally of good but
simple appearance.
When provided with spaced elastomeric bearing pads at their tops, as
described in Section 5.5, hollow prismatic piers of adequate proportions
are especially functional in structures that are erected in cantilever. This is
partially because their capability of resisting bending moments negates the
164 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

Fig. 5.13 A pier formed of twin vertical walls.

need of temporary struts during erection. In addition, as explained in

Section 6.7, the provision of elastomeric bearing pads between superstruc-
ture and substructure provides a means of adjusting the structure during
The structural analysis of piers which incorporate elastomeric bearing
pads has been described by Mathivat (Ref. 60) who has shown the relation-
ships between the rotation 8, horizontal translation 6, vertical deflection v
and the moment M, shear Q and vertical load N applied at the top of the
pier can be written as follows:

E8 = AM + BQ (5.17)

E6 = BM + CQ (5.18)

Ev= KN (5.19)

in which for a pier alone (no elastomeric bearing pads) the elastic constants
for rotation (A due to moment and B due to shear), displacement due to
shear (C) and vertical deflection (K)* are as follows:

*Note: B is both the coefficient for rotation due to shear and the coefficient
for displacement due to moment. This can be explained by Maxwell’s law.

h dx (5.20)
A, =

BP= - (5.21)
s0 I

x ’ d x
- (5.22)
s0 1

K, = h dx
s“ A

Fig. 5.14 A flared pier.


G = shear modulus of elastomer

p = number of pads per row
n = number of laminations in each pad
S = a x b (area of each pad)
t = thickness of each lamination

tr a




Cn=hBp-Cp Cn>hBp-Cp Cn=o0

Fig. 5.15 Pier with spaced elastomeric bearing pads.


For a prismatic pier, Eqs. 5.20 through 5.23 become:

A,, = + (5.24)

BP = E (5.25)

c, = g (5.26)

in which I and A are the moment of inertia and area of the pier cross section
respectively and h is the height of the pier. If spaced elastomeric pads,
having the dimensions shown in Fig. 5.15 are placed on top of a pier the
values of the elastic constants A, B, C and K in Eqs. 5.17, 5.18 and 5.19
must include the effect of the bearing pads. These effects are:

A, cc ;z*, t3 (5.28)

B, = 0 (5.29)



in which all terms except c are defined in Fig. 5.15.

TABLE 5 . 2

b/a 0.5 0.6 0.7 0.75 0.8 0.9 1.0 1.2 1.4 1.5 2 3 4 5 10 30
c 5.83 4.44 3.59 3 . 2 8 3 . 0 3 2 . 6 5 2.37 2.01 1 . 7 8 1 . 7 0 1 . 4 6 1 . 2 7 1 . 1 8 1 . 1 5 1 . 0 7 1

The coeffkient c has the value shown in Table 5.2. It can be shown that for
the combined effects of the pier and elastomeric pads, the elastic coeffi-
cients are as follows:
168 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

A = A, + A, E (5.32)

B = B, (5.33)
C = Cp + C, E (5.34)

K = Kr + K, E (5.35)

Using the above, it can be shown that for a pier which has no horizontal
deflection (6 = 0) the moment at the base, Mb can be determined from

Mb = (I - C T’; E) M (5.36)
P e
in which M is the moment at the top of the pier. The term
, - bh (5.37)
cp + CeE
is the carry-over factor. Consideration of Eq. 5.37 will reveal that for the
case of a prismatic pier without elastomeric bearing pads the value of Eq.
5.37 is -0.50 (the familiar carry-over factor). On the other hand if
t C,E = hBF - C, (5.38)
the carry-over factor is equal to zero and there is no moment at the bottom
of the pier. Finally, as the value of Ce approaches infinity, the carry-over
factor approaches unity. These are illustrated in Fig. 5.15.
Consideration of these fundamental elastic principles will reveal the
important advantages that can be realized through the use of spaced elas-
tomeric bearings. Their use permits the engineer to control the elasticity of
the substructure and transfer bending moments due to vertical loads bet-
ween superstructure and substructure. In addition, the effects of length
changes can be controlled.

Piers having twin walls. Rather than employing spaced elastomeric bear-
ing pads as shown in Fig. 5.7 to control the effects of volume changes in
bridges having piers of small to medium height, piers having twin vertical or
I inclined walls as shown in Figs. 5.13 and 5.16 have been used. This pier
configuration behaves in a manner similar to that of a hollow prismatic pier
with spaced elastomeric bearing pads (Ref. 60). The walls may be fixed to
the superstructure as well as to the solid base of the pier. Piers have also
been made with concrete hinges at one or both ends of the walls. Longitud-
inal length changes such as those caused by temperature variations, shrin-

Fig. 5.16 A pier formed of twin inclined walls.

kage and creep cause bending of the wall elements which in turn results in
very little resistence to these movements. On the other hand, under the
action of vertical loads, the structure containing twin wall piers behaves as
a frame and moment is transferred from the superstructure to the substruc-
ture. Unbalanced superstructure moments can result in uplift forces being
applied to one of the two walls of a pier and if the uplift force exceeds the
dead load in the pier, the wall will act as a tension member and must be
reinforced accordingly (Ref. 61). When converging walls are used, a por-
tion of the longitudinal deformation of the superstructure is resisted by
truss action as contrasted to bending alone in the case of vertical walls. If
converging walls which have hinges at both their tops and bottoms are used
on slopes that result in their axes converging at the bottom of the founda-
tion, as shown in Fig. 5.16, the foundation is subjected to vertical load
alone. On the other hand, if the axes of the walls converge at a point above
or below the bottom of the foundation, the foundation must resist a bending
moment in addition to the vertical load. Needless to say, thin pier walls of
this type must be investigated for elastic stability, This type of pier is
relatively economical in segmental construction because they are stable
during construction with very little in the way of temporary bracing. The
appearance of piers of this type can be quite acceptable as shown in Figs.
5.17 and5.18. Adetailed study ofpiersofthis type iscontained in Ref. 60.
Fig. 5.17 Choisy-Le-Roi Bridge, France (Courtesy of Sock%6 Technique pour L’Utilization de
le Prkontrainte).
Fig. 5.18 Chillon Viaduct, Switzerland. (Courtesy of SociQ6 Technique pour L’Utilization de
la Prbcontrainte).

Flared piers. Piers which are provided with flared upper portions, such
as shown in Figs. 5.19 and 5.20 have been used on a number of bridges.
Piers of this shape are not normally used on structures which are low (short
piers). The pier shafts may be solid but they are usually hollow because the
shafts themselves generally are quite large in cross sections. The piers may
have openings, recessed areas or “special architectural details” which are
intended to enhance their appearance. The flared outer edges of the piers
near their tops as well as the openings, recesses and other architectural
details often complicate the structural details of the piers. This in turn
results in structural inefficiency, imposition of specific construction
methods that must be used in the pier and superstructure and additional
cost. The flares at the tops of piers are particularly a problem in precast
segmental construction when the design is predicated upon a fixed connec-
tion between superstructure and substructure. The construction details
required to effect such aconnection are complicated and costly. The use of
piers with flared tops is less objectionable with cast-in-place segmental
construction. Many persons feel piers of these types are attractive and
Fig. 5.19 Mission Valley Viaduct, San Diego, California. (Courtesy of State Department of

Fig. 5.20 Pine Valley Creek Bridge, California. (Courtesy of State Department of Transporta-

hence the additional cost they entail is worthwhile. Others feel the basic
shape of the pier should be chosen to satisfy its functional (structural)
requirements with due consideration to cost and appearance.

5.7 Seismic Forces

The earthquake design provisions of the AASHTO Specification are con-

tained in the Interim Specifications issued in 1975 (Ref. 62). These provi-
sions are an outgrowth of the significant amount of bridge damage that
occurred in the 1971 San Fernando earthquake. It was this event that
prompted the California Department of Transportation to re-evaluate the
seismic design procedures then being used as well as to re-examine details
of construction which were being used as standard details (Ref. 64). This
study resulted in the seismic design criteria that are currently used in
California and which are contained in the AASHTO Interim Specitica-
tions (Ref. 63). The present criteria requires consideration of the relation-
ship of the site to active faults, soil-structure interaction, and the dynamic
response characteristics of the bridge as a whole. The criteria is among the
more sophisticated in use today.
6 Construction

6.1 Introduction

Many factors affect the relative cost of the various modes of bridge con-
struction. The relative importance of the factors varies throughout the
country (and the world) and hence escape generalization. For this
reason no attempt has been made to give specific cost data nor even
relative cost data for the different bridge types. Construction considera-
tions vary in importance between the bridge types and an attempt is
made to discuss the basic considerations of falsework, formwork, con-
crete finishing and camber control.
Girder and box-girder bridges have had wide use in the United States
for a number of years and it is doubtful if major innovations will evolve
with these modes of construct4on. Segmental bridge construction has not
as yet been widely used in the United States, in spite of the fact that
several hundred segmental bridges have been constructed in other parts
of the world. It is expected that many innovations will be made in this
country when the use of this method becomes more common.
1 7 6 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

6.2 Falsework
Bridges which are cast-in-place on falsework are normally of the type
that are classified as T-beam or box-girder bridges. Bridges have, how-
ever, been constructed on falsework, even though they have been de-
signed to be erected segmentally in cantilever. This is illustrated in Fig.
6.1. This permits advantage to be taken of the economies of both
methods. The falsework normally consists of timber or steel beams sup-
ported by timber or steel columns. A typical bent which utilizes steel
beams supported on timber posts is shown in Fig. 6.2(a). Typical details
for falsework composed of steel beams supported by steel towers are
shown in Fig. 6.2(b).
In recent years the trend has been toward greater use of steel
falsework and less use of timber falsework. This has partially been the
result of improved modular systems of steel falsework becoming com-
mercially available and partially due to timber becoming more expensive
than it was in previous years.
The costs related to falsework can represent a major portion of the
total cost of a bridge. The falsework related costs, which are in addition
to the purchase or rental of the basic falsework materials, include the
1 . Engineering costs for design and inspection of the falsework.
2 . Preparation of the foundation. This might consist of preparing soil


i--J L--j




Fig. 6.1 Falsework for use with the segmental erection technique.

‘,L-----A L-----J

Fig. 6.2 (a) Typical timber falsework

c ----I I - - - - - ; [-----I r----7 r-----7 r-3

‘,‘; I i
\\ / I 1I I 1 1:
\ ‘1 II I I I i I I /I
\L----ii-----J L-----A L-----J L---j/
I ,

Fig. 6.2
(b) Falsework incorporating steel beams and steel towers

and setting wood mudsills, concrete pads or perhaps driving tem-

porary piling. In many cases wood mudsills can be used on level,
fine graded natural soil or thin blankets of sand placed over
natural soil. In some locations the natural soil has low strength
and it is removed and replaced with well-compacted soil that is
adequate for the loads that are to be imposed by the mudsills or
concrete pads. Temporary piling are used in cases where it is not
feasible to use mudsills.
3 . Fabrication or assembly and erection of the falsework bents, to-
gether with adequate longitudinal and transverse bracing.
4 . Setting the beams, whether of wood or steel, which span between
falsework bents.
5 . Fabricating and setting camber strips on the beams, if required by
anticipated beam deflection. (This is often done before the beams
are erected.)
6 . Adjusting the assembled falsework to grade. (Generally done after
the formwork has been partially assembled on the falsework.)
7 . After the concrete has been placed, cured and stripped of the
formwork, lowering the falsework, removal of the beams and
striking the falsework bents.
8 . Removing the mudsills or temporary piles and restoring the site.
This may include reconstructing abutment slopes in areas where
falsework bents must be set on the slopes.
Falsework which must be erected over an existing highway or railroad
is considerably more difficult to erect if traffic must be maintained. It
also involves considerably more risk to the contractor. Furthermore,
some agencies require especially conservative designs and details for
falsework which is to be erected over a traveled way. For these reasons,
falsework costs are greater under such circumstances.
Bodies of water, mudflats and deep valleys can add materially to the
problems associated with the construction of falsework. Costs are also
affected by obstacles of these types.
Falsework, being a very temporary structure, is normally constructed
using materials, connections and details which would not be permitted
or even considered in a permanent structure. Because of this, falsework
is more subject to failure, whether originating by accident, or inadequacy of
materials, connections and details, than are permanent structures. This
fact is reflected in the cost of structures erected on falsework.
Bridge superstructures which utilize precast girders or segmental gir-
ders which are erected in cantilever, frequently do not require any
falsework. As was explained in Section 4.2, falsework at ,the piers of
segmental bridges is sometimes needed to accommodate the unbalanced

moments which exist during erection. Falsework of this type is quite

special and requires careful consideration in its design in order to
achieve adequate strength together with details which facilitate its erec-
tion and removal.

6 . 3 Formwork

The cost of formwork, per unit of area, varies significantly between the
various modes of prestressed concrete bridge construction. Surprisingly,
the unit price (cost per square foot of bridge) of bridge types which re-
quire more jobsite constructed formwork may be lower than bridges of
other types. The phenomenon is attributable to the quality of the
formwork required together with the degree of prefabrication that is pos-
sible and the working conditions available for the assembly of non-
.prefabricated forms.
Box-girder bridges of normal configuration have more formwork con-
tact area per unit area of bridge deck than T-beam or precast prestressed
concrete girder bridges. On the other hand, the formwork for box-girder
bridges is relatively easy to construct and much of it does not need to be
of high quality material nor fabricated with high precision. The sofftt
formwork generally consists of 4 x 4 wood sleepers placed transversely
on top of the falsework beam camber strips over which 5/8-inch plywood
is nailed (nominally). The sofftt fonnwork is easily constructed because
there is little (if any) cutting and fitting of material and the work can all
be done “down-hand”. High quality formwork is required for the deck
slab overhang and the exterior surface of the fascia girder. The surfaces
of the interior webs do not require high quality formwork because they
are hidden from view in the completed structure. Formwork for the in-
terior webs is usually reasonably well made with material of good quality
because they are frequently reused many times. The formwork for the
upper slab of the interior cells of box-girder bridges is normally not
stripped after the slab has cured and for this reason it is frequently refer-
red to as the “lost-deck-form”. Because the lost-deck-form is used only
once and because the surface it forms is not subject to view in the com-
pleted structure, material of very low quality together with only
adequate workmanship is normally used in its construction., Stripping of
all the formwork (webs, slab overhangs and soffit) is relatively easy and
little must be done “over-head”. This combination of requirements re-
sults in relatively inexpensive formwork and even though the quantity
which is required is relatively large, the overall formwork cost is rela-
tively low.
180 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

The quantity of formwork required for cast-in-place T-beam girder

bridges is frequently less per unit of deck area than is required for box-
girder bridges. The soffrt formwork is relatively expensive because it
must be made to a specific width, must be vertically cambered and made
horizontally straight. The side forms for the diaphragms and all the
webs, both interior and exterior, must be well constructed of high qual-
ity material in order to obtain straight smooth lines and surfaces. The
same is true of the deck formwork. The forms do lend themselves to a
reasonable amount of prefabrication. Stripping T-beam forms is difficult
because at least the deck form must be done “over-head”. All surfaces
of a T-beam bridge are visible in the completed structure.
Although the amount of field assembled formwork required for
bridges which incorporate precast prestressed concrete girders is rela-
tively small, it is also rather difficult and costly to construct. The in-
terior diaphragms are normally relatively thin and the girders must be
tied together when the diaphragm concrete is placed in order to prevent
the pressure of the plastic concrete from deflecting the girders laterally.
The interior diaphragm forms must be field cut and adjusted to fit the
girders as built and as erected and hence do not lend themselves to com-
plete prefabrication. Little exists in the way of safe or convenient work-
ing space for the construction of the diaphragms. They are normally
constructed (and stripped) before the deck is formed. The deck forms
are frequently supported from adjustable metal brackets as shown in
Fig. 6.3. Because the space between the girders is never exactly straight
nor with the precise width shown on the construction drawings, the deck

Fig. 6.3 Forms for the deck of a girder bridge incorporating precast girders.
forms must be field cut to fit the actual space between girders. The rein-
forcing steel stirrups which project from the tops of the girders hinder
the construction of the deck form. Strippmg the deck forms is always
over-head work and frequently must be done under adverse and poten-
tially dangerous conditions. Formwork quality must be high because the
surfaces are visible in the completed structure and leakage during plac-
ing of the concrete must be minimized as is explained below in Section
6.4. These conditions result in relatively costly formwork.
The traveling forms, or simply the travelers, which are used in cast-
in-place segmental bridge construction perform two functions. These are
the forming of the desired shape of the concrete segments and holding
the forms, falsework, concrete and other materials in the desired posi-
tion, cantilevered from the previously constructed work. These func-
tions may be done by two independent portions of the travelers or may
be done by a single device which performs both functions. The
formwork function of the travelers is normally done with steel forms
although particularly in the vicinity of the piers, where transverse dia-

Fig. 6.4 Traveling form for cast-in-place segmental bridge construction. (Courtesy of
California Department of Transportation).

phragms must generally be provided, the internal formwork may be done

with wood. The structural cantilevering function of the travelers is nor-
mally done with steel trusses which incorporate hydraulic equipment for
advancing and adjustment for line, grade and superelevation. When the
traveler consists of a single device, the side forms act as structural gir-
ders cantilevered from the previously constructed work. Needless to
say, the cost of the travelers is great and must be amortized over a
number of reuses. When used in erecting bridges over deep valleys,
dismantling the traveler after completing one cantilever and moving it to
another bent to begin a new cantilever can be a somewhat difficult and
time consuming task. A typical traveler is shown in Fig. 6.4.
The equipment and method of casting precast segments for segmental
bridge construction has gone through three generations. Innovations in
these methods will certainly come about in the future. The method first
used is shown in Fig. 6.5, and is sometimes referred to as the “long-line
method”. In this, the simplest of the methods, the segments are cast on
a fixed concrete soffit form, each segment being cast against the adja-
cent one, with the internal and external forms moving down the soflit
form as the construction progresses. The procedure of casting one seg-
ment against a previously cast segment is referred to as match-casting
and the segments are referred to as match-cast segments. The forms are
generally of steel but they can be made of wood. The segments are re-
moved to the bridge site and reassembled in the same relative position in
which they were cast. A variation on this technique which was
employed on the bridge between the European Continent and the Island
of Oliron in France is shown in Fig. 6.6. This variation permits the start



Fig. 6.5 Principle of the precasting technique for segmental box girder. construction
utilizing a long soffit form.







Fig. 6.6 Principle utilized in precasting the segments for the bridge between the continent
and the Island of Oleron.
184 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S








Fig. 6.7 Casting cell for precasting segments.


of precasting the segments for a second cantilever before the completion

of the first. This is accomplished by the provision of a means of remov-
ing the pier segment before the first cantilever has been completed. The
cost of the concrete soffit forms is relatively great and the method is
only considered practical on bridges which have constant horizoptal and
vertical curvature (if any at all). The second generation of precasting
segments involves the use of a casting machine or casting cell, and can
accommodate variable horizontal and vertical curvature as well as vari-
able superelevation. The method is sometimes referred to as the “short-
line method”. The principle of the casting cell is illustrated in Fig. 6.7.
It consists of adjustable internal and external forms which are always
used in a level and plumb position. The previously cast segment is set
next to the form in the proper relative position with respect to the seg-
ment being made to give the required relative position in the final struc-
ture. The new segment is cast against the older segment insuring perfect
fit when later it is being reassembled. The cost of the casting cells used
in the second method varies between wide limits but is high and necessi-
tates amortization over a number of reuses. The third precasting
technique consists of match casting the segments in a vertical position,
the new segment being cast on top of the previously cast segment as
shown in Fig. 6.8. The method can be thought of as the “vertical short-
line method”. The basic steel form system is somewhat like that which

3 n

- -

2 n-l

First Second Third nth

step Step Step Step

. Fig. 6.6 The technique used in match casting segments in a vertical position.
___-_ x

Fig. 6.9 Floating crane erecting a precast girder. (Courtesy of Freyssinet

Company, Inc. New York).

Fig. 6.10 Launcher for erecting precast girders.


is used in the manufacture of concrete pipe. The third method has been
used exclusively on the production of a large number of freeway over-
crossing structures in southeastern France and is particularly useful in
producing segments which have small depths.

6.4 Concrete Finishes

Specifications for the finishing of concrete bridges vary between the var-
ious jurisdictions. Hence, the statements made herein in this regard,
which are based upon the standard specifications used in highway bridge
construction in California (Ref. 63, may not be universally applicable.
In general two types of finishes are required. These are termed ordi-
nary surface finish and Class 1 surface finish. Ordinary surface finish
consists of filling all holes or depressions in the surface of the concrete,
repairing rock pockets and repairing honeycomb as well as removing
fins, stains and discolorations which are visible from traveled ways. Or-
dinary surface finish is the final finish for most surfaces on a bridge and
is a preparatory finish to surfaces required to have a Class 1 surface
finish. The Class 1 surface finish includes eliminating any bulges or de-
pressions or other imperfections which exist due to the forms not being
of high quality. The removal of bulges requires grinding the surface after
which they must be “sacked” to give a uniform appearance.
The Class 1 surface finish is required for the deck overhang and ex-
terior faces of the exterior beams or webs of all bridge types. Hence the
cost of this portion of the work is roughly the same for bridges of all
The inside surfaces of box-girder bridges do not require fin, stain or
discoloration removal. Ordinary finish is required for the deck, and
beam sofftts as well as the sides of the beams in girder bridges. Only the
bottom deck surface of a box-girder bridge requires the ordinary finish.
In the construction of girder bridges in which the deck is not placed
monolithically with the girders, care must be exercised during the plac-
ing of the deck concrete to insure that concrete does not leak through
the deck-form-to-girder joint and become deposited upon the side of the
girder. If such deposits do occur, they should be washed off with fresh
water before they harden. Otherwise they increase the effort (and cost)
required to properly finish the concrete surfaces.
As a result of the above, concrete finishing costs are usually less for a
box-girder bridge than for girder bridges. Segmental bridges should have
finishing costs comparable to those of a box-girder bridge.
188 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

6.5 Erection of Precast Girders

The erection of precast girders to be incorporated in structures crossing

over an existing roadway is normally accomplished with the familiar truck
crane. For very large girders or for larger than usual reaches, crawler
cranes may be used. Girders which are to be used in bridges which cross
bodies of calm water are frequently erected with floating cranes as shown
in Fig. 6.9. The erection of precast girders under these circumstances is
straightforward and does not require further explanation.
For precast girders to be used in bridges crossing deep valleys or bodies
of water subject to sudden changes in conditions (such as a river in the
tropics) or those having strong and variable current, the use of a girder
launcher may be the best approach. A girder launcher is shown in Fig. 6.10.
In using a girder launcher the launcher is first cantilevered over the first
span until it can be supported near the first abutment and the first pier. The
girders for the first span are then picked up by the launcher and moved into
position between the first abutment and the first pier. Some launchers are
equipped to move the girders transversely while others are not. When the
launchers are not equipped to move the girders transversely, they are
moved horizontally by placing them on greased steel plates and jacking
them laterally. Before a launcher can be advanced, it is frequently neces-
sary to install the permanent or temporary diaphragms between the erected
girders in the first span, placing a lateral bracing system in the plane of the
top flanges of the girders to prevent buckling of the compression flanges
and finally placing a temporary wood deck on the first span to provide a
surface over which the launcher and girders can be moved without interfer-
ence from the bracing system and reinforcing steel dowels which usually
protrude from the tops of the erected girders. The launcher is then ad-
vanced and the girders in the second span are erected and the procedure is
repeated until the bridge is erected from one abutment to the other. It is
generally not possible from a structural viewpoint, nor feasible from a
construction scheduling viewpoint, to cast the permanent deck slab in
place and thereby avoid the temporary top flange bracing and wood deck.

6.6 Erection of Precast Segments

Precast segments have been erected using a variety of methods. In select-

ing an erection method one must consider three major factors; the cost
of the equipment required in the method, the rate at which the equip-
ment will erect segments, and the constraints imposed by the site. Vir-
tually all large structures have been erected with a gantry designed ex-
pressly for the particular structure. The specially designed gantries are
the preferred method because they are capable of erecting the segments
more rapidly than any other equipment and they can be used regardless
of site conditions. The relatively high first cost of the gantries preclude
their use on less important structures.
The simplest method of erecting precast segments, and the one which
involves the least capital investment, is to employ a conventional truck
or crawler crane. Cranes of these types are readily available in virtually
all parts of the country and most bridge contractors are familiar with
their use. There are, however, several reasons why the use of conven-
tional cranes may not be feasible on a particular structure. The load
capacity of a crane is a function of the size of the crane, the length of
boom required to reach the required height and the load radius (mea- .
sured from the axis about which the crane cab revolves). The capacity
of a crane reduces rapidly as the height of lil’t and load radius increases.
Some combinations of segment weight, pier height and minimum possi-
ble load radius preclude the use of even the largest conventional crane.
It is normally not feasible to “walk” a crane which is under a large load.
Hence, the crane must be positioned directly below the final position of

Fig. 6.11 Erection of the Choisy-Le-Roi Bridge with floating cranes. (Courtesy of Sock% pour
L’Utilization de le Prkontrainte, Paris, France).
190 ( H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

the erected segment and the segment must be transported to the crane.
Site restrictions may preclude such access. It should also be recognized
that conventional cranes are relatively flexible, particularly with the
longer booms, and because of oscillations and the difficulty of raising or
lowering a load slowly in minute increments, segment erection with con-
ventional cranes is generally slow. Segment erection requires greater
precision in positioning the load being erected than is required in the
erection of precast beams.
Bridges which traverse bodies of water can often be erected with fldat-
ing cranes as shown in Fig. 6. Il. Floating cranes of very large capacity
exist, and very long booms can be used with them where necessary. The
effects of wind, current, waves, tide and wakes of moving vessels can
result in conditions which render the use of floating cranes difficult and
slow, if not impossible. One solution for this problem is to use the float-
ing crane to erect the segment on a secondary piece of erection equip-
ment that is supported on the bridge superstructure, the secondary
equipment being used to adjust the segment into final position. Secon-
dary equipment of this type was used in the bridge shown in Fig. 6.11. It
would be practical for use with conventional truck or crawler cranes.
When it is possible to transport the segments to a position directly
below their final position in the completed structure, equipment of the
type shown in Fig. 6.12 can be used to erect the segments. This type of
equipment is considerably less costly than a gantry but is also much
An erection gantry was used for the first time in precast segmental
bridge construction in the Oliron Bridge. This gantry is shown in Fig.
6.13. There have been a number of bridges constructed by gantry since
the construction of the Ole’ron Bridge. The design of the gantries has
gone through a constant evolution and their efficiency has been progres-
sively increased. As an example, the gantries used in the construction of
the Rio-Niteroi Bridge in Brazil were able to erect a typical span of 262
feet every seven days. The Rio-Niteroi Bridge is shown during construc-
tion in Fig. 6.14. Gantries can be used to erect segments which are
transported to the gantry over the top of the previously erected
superstructure as was the case with Oleron or which are transported to a
point below the gantry as was the case with the Rio-Niteroi. In in-
stances where the pier or pier-superstructure connection is not sufli-
ciently strong to resist the unbalanced moment induced by the cantilever
erection technique, the magnitude of the unbalanced moment can be
minimized by erecting two segments simultaneously as shown in Fig.
6.15 or its effect can be nullified by using the gantry itself to brace the
structure during erection as shown in Fig. 6.16. Erection gantries are
- Remove
truss after

L Step 2 Step 3

Fig. 6.12 Erection procedure with segments transported to a position directly below their
final position.

Fig. 6.13 Gantry erecting precast segments on the Ol&on Bridge. (Courtesy of Soci&~ pour
L’Utilization de la Pr&xntrainte, Paris, France).
Fig. 6.14 Rio-Niieroi Bridge during construction. (Courtesy of !Soci&h pour L’lJtilization de la
Prbcontrainte, Paris, France).

very expensive but no other erection method is as versatile or as fast.

The erection of bridge superstructures is normally done by starting at
one end of the structure and working toward the other. There are many
reasons for this. In the case of erecting precast segmental bridges with a
gantry, it is especially important to erect the superstructure in this man-
ner. One of the reasons for this involves the stability of the structure due
to its own dead load and is illustrated in Fig. 6.17. It will be seen in Fig.

Fig. 6.15 Gantry erecting two segments simultaneously.


the cantilevers

Fig. 6.16 Gantry used to brace the superstructure as well as to erect the segments.

6.17 that the structure is rendered stable after each new cantilevered
section has been erected. The stability is achieved by connecting the
newly erected cantilever to the previously erected portion of the
superstructure. In the case of the first cantilever, the stability is
achieved by completing the first span. If all the cantilevers were to be
erected before continuity was established between the cantilevered por-
tions of the bridge, the construction cost and risk would be greater. This


I 1
t n

LI r’

t r’


Fig. 6.17 Recommended erection sequence with gantry+rected precast segments.


1 n

r I I
7 r

‘I 1

Fig. 6.18 Erection sequence not recommended with gantry-erected precast segments.

procedure is illustrated in Fig. 6.18. The procedure shown in Fig. 6.18

involves a more elaborate and costly gantry than is required when con-
tinuity is established as the erection progresses. This is because the gan-
try must advance without imposing any load on the cantilevers.
An erection method which utilizes a temporary cable-stayed technique
was used in the summer of 1974 for the first time on the Rombas Bridge
in France (Ref. 66). This method, which is illustrated in Fig. 6.19, per-
mits the bridge to be erected from the top with the segments being
transported to the point of erection over the top of the previously
erected superstructure. It has two major disadvantages; (1) the first end
span must be erected on falsework or with some other special method
and, (2) the bridge substructure is subjected to larger dead loads during
erection than in the completed structure and hence must be designed
therefor. The cost of the equipment with this erection method is modest
and the speed of erection is not as fast as with a gantry. Considerable
engineering effort is required to determine the adjustments required in
the loads in the temporary cables during the progression of the bridge
A large number of small grade separation structures were constructed
using precast segments on the Rhone-Alps Motorway in France. The
structures have central spans up to 100 feet long. The erection of the
segments was accomplished with conventional truck cranes together.
with a minimum amount of steel falsework and a clever temporary pre-
stressing scheme. The erection procedure is illustrated in Fig. 6.20 in
C O N S T R U C T I O N C O N S I D E R A T I O N S 11 9 5

which it will be seen that the first segment is erected on top of four
25ton hydraulic jacks after which the second and third segments are
connected to the first with temporary prestressing. The temporary pre-
stressing which connects the first three segments includes top and bot-
tom prestressing. Four lOO-ton jacks, which are equipped with low-
friction slide plates, are then installed to support the segments and the
four 25ton jacks .are removed. The segment erection is continued using
the cantilever technique with the prestressing being provided by tempor-
ary bar tendons attached to the tops of the segments. After each half of
the segments have been erected, the assemblies of segments are pushed
together thereby eliminating the need of a cast-in-place closure joint.
The erection procedure is completed by installing the permanent pre-
stressing tendons (in two stages), erecting the abutment segments, and
removing the temporary prestressing tendons as well as the falsework.
Superstructures can be erected in as little as four days with this

6.7 Camber Control

Prestressed concrete structures deform under the effect of short dura-

tion loads in much the same manner as structures composed of other
materials. Because concrete is a material which is subject to time-
dependent deformations as a result of creep and shrinkage, and because
different prestressing steels exhibit relaxation to different degrees, the
long-term deformation and deflection of prestressed concrete members
under dead load alone is the result of a complex interaction of several

Fig. 6.19 Erection scheme used on the Rombas Bridge in France.


Temporary Tie Bars

Move assemblies to
close joint

Install first stage

/- penenent tendona

,,- Erect end and abutment segmenta -.,

secondstage permanentpresttussing
and remove tie by.

Fig. 6.20 Erection technique employed on the RhBne-Alps Motorway overcrossing


phenomena. In order to achieve desirable results during the service life

of a structure, the structure must be constructed with the effects of
time-dependent deformation being taken into account. Considerable in-
formation is available in the literature regarding the computation of
long-term deformation of prestressed concrete members and hence the
fundamentals of this subject are not repeated here.
The effect of time-dependent deformations on the deflection of cast-
in-place prestressed concrete structures is taken into account by building
camber in the formwork used to construct the structures. This is a well
known procedure which does not require further explanation. In a simi-
lar manner provision is made for camber in structures utilizing precast
girders through the use of a haunch of variable depth over the girders as
was described in Section 2.7. No further consideration of this subject is
The estimation of the deflection of bridges which are constructed
segmentally is more complex than for more conventional prestressed
concrete structures. The additional complexity partially emanates from
the fact that the age of the concrete varies from segment to segment and
hence the elastic modulus is variable along the length of a segmentally
constructed member. In addition, because the tendons are prestressed at
different time intervals, the various tendons are relaxing on a different
time basis. When the cantilever erection technique is employed, the de-
flections must be computed for each step in the construction. The most
logical approach to analyzing a structure with so many time-dependent
variables is to use a numerical integration procedure in which the many
variables can be treated as independent time functions.
In the case of cast-in-place segmental bridges, from four to seven days
will normally be required for a cycle in the fabrication of segments of
usual size. For larger segments greater periods are required. Because a
cast-in-place cantilever may have from six to twenty segments, it will be
realized that the difference in age between the first and last segments
cast in any one cantilever might be as great as 130 days. Using the
strength and elastic modulus relationships proposed by Committee 209
of the American Concrete Institute which are:

fct = (-&) f, (6.1)

196 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

In which
f’,t = concrete strength at the age of time t
f, = concrete strength at the age of 28 days
c,d = coefficients for concrete strength which are taken to be 2.24 and
0.92 respectively for water cured concrete of usual quality.
(Type III Cement).
Ect = elastic modulus of the concrete at the age oft days
W = unit weight of the concrete which is taken as 145 pcf for normal

one can show that in any one cantilever the elastic modulus can vary
from 3,660,OOO psi to 4,210,OOO psi. The oldest concrete has an elastic
modulus equal to 115% of that of the youngest. Obviously variations in
elastic properties as great as this should be taken into account in com-
puting the deflections of the cantilevers. Furthermore, in the case of
cast-in-place construction, because the concrete is prestressed and in-
corporated in the structure at a relatively early age, a high portion of the
total concrete shrinkage is effective in reducing the prestressing force.
Additionally the total creep strain is much greater for concretes which
are stressed at early ages. This is illustrated graphically in Fig. A.6 in
Appendix A where it will be seen that concrete stressed at the age of 3
days exhibits a creep strain that is 160 percent of that for the same con-
crete being stressed at an age of 28 days, all other conditions being
equal. This effect also has a tendency to increase the loss of prestress,
add to deflection and render the accurate prediction of the deflection
more difftcult.
In the case of bridges composed of precast segments, the segments of
any one cantilever are frequently cast one per day with the result that a
cantilever which has 20 segments may have a difference in age between
the oldest and youngest segment of the order of 25 to 28 days (including
an allowance for weekends). Because the youngest segments would
normally be 30 days old or more at the time of erection, the ratio of the
elastic modulii of the oldest and youngest concrete at the time of erec-
tion would be of the order of 1.02 or less which is negligible. Because of
the rapid rate of erection normally achieved with this mode of construc-
tion, the amount of creep which takes place during erection is less than
that experienced with cast-in-place segments and the total creep strain is
less because the concrete has a greater age at the time of stressing. For
these reasons the loss of prestress is relatively low with precast seg-
ments as is the total deflection of the cantilever at the time of erection.
Studies of the relative deflection of cantilevers composed of cast-in-
place and precast segments have been reported in the literature. (Ref.
IAINT fl 1 2 3 4 5


JOINT 0 1 2 3 4 5

Fig. 6.21 (a) Deflection of a segmentally constructed cantilever in which no provision has
been made for camber. (b) Deflection of a segmentally constructed cantilever in
which provision is made for camber.
I 2 0 0 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

67.) These studies have shown the deflection of the former can be as
great as 2.5 times that of the latter for identical cantilevers constructed
under normal conditions for each construction mode.
Deflection control for a bridge erected in cantilever involves the de-
termination of the deflection of each cantilever as it is assembled fol-
lowed by the determination of the deflection of the total structure as new
portions are added to it. These are illustrated in Figs. 6.21 and 6.22. A
typical deflection curve for a cantilever that is erected segmentally with-
out camber but including the effects of prestressing is shown in Fig.
6.21(a). The deflection curves for each stage of construction for the
same cantilever in which the appropriate correction for camber has been
built into the member, is shown in Fig. 6.21(b). The cantilever camber
that is to be constructed into the structure at any point, such as point 3
in Fig. 6.21(a), is the deflection the point undergoes from the sub-
sequently erected segments. In other words, point 3 must be constructed
with a camber equal to the sum of the deflections due to the effects of
segments 4 and 5. These are shown graphically in Fig. 6.21 in which
they are designated 83.4 and 63.5. The deflection which must be taken
into account in the construction of a four-span continuous bridge that is
erected in cantilever, in addition to the cantilever deflection, is shown in
Fig. 6.22. The downward deflection of the cantilevered ends results
from the stressing of the continuity tendons in each span as the con-
struction proceeds. The downward deflection of the cantilever accounts
for the discontinuities in the deflection curves.
In cast-in-place construction the deflection analysis must also include
the effects of the traveling form because it contributes significantly to
the deflection of the cantilever. Because the traveler is moved forward
as the construction progresses, the amount of deflection attributable to
the traveling form progressively increases. When the construction of the
cantilever is completed, the traveling form is removed and the elastic
deflection caused by the traveling form is recovered.
The camber is built into the structure in cast-in-place cantilever con-
struction much in the same way that it is in conventional cast-in-place
construction. That is to say the elevation of the traveling form is field
adjusted by a crew composed of surveyors and workmen in such a way
that the desired camber is built into the structure. In addition, as the
construction progresses, the surveyors periodically measure and plot the
camber in order to compare the actual values to the computed values. If
it is found the actual values differ significantly from the computed val-
ues, the camber provided in subsequently constructed segments can be
adjusted to correct for the deviation between the two values.
In the case of precast segments which are provided with relatively

,I 4,
SPAN 2 SPAN 3-- -SPAN 4 .

- I

/ (d) Completion of Spans 1, 2 and 3 includi

I prestressing of Spans 1, 2 and 3. I

(e) Completion of Spans 1, 2, 3 and 4 including continuity

prestressing of Spans 1, 2, 3 and 4.

tendon extending from abut. to Abut. 5.

(g) Complqted structure including superimposed dead load.


Fig. 6.22 Deflections occurring during construction of a four-span bridge erected in

cantilever, in addition to the cantilever deflection.

thick joints of mortar or concrete between the segments (a procedure

which is not normally employed in contemporary segmental bridge con-
struction),camber control can be handled with the precast segments in
much the same manner as described above for cast-in-place construc-
Precast segments which are match cast are normally erected with thin
joints which are of the order of 0.02 inches thick. Hence the camber
must be provided in the shape of the segments themselves. To ac-
complish this when the segments are precast on rigid soffrt forms (long-
line method), the camber is built into the soffit form in much the same
way as is done with conventional cast-in-place construction. When cast-
ing cells of either the horizontal or vertical type are used, the camber
adjustment is made by orienting the previously cast segment, with re-
spect to the segment in fabrication, in such a way as to provide the
required relative position between the two segments. Horizontal and
vertical Curvature as well as superelevation, if any, are provided for in
the same manner. The adjustment of the previously cast segment with
respect to the one under fabrication is done under the direction of a
surveying crew working in the precasting yard. After each new segment
is cast, the segments are again measured before they are removed from
the casting cell and the data plotted as a means of insuring the work is

IO ! 01

Hydraulic Jacks
Elastromeric bearing pads ’
Vertical Section Through
Bridge at Pier

Fig. 6.23 Procedure for adjusting the final position of a cantilever.

C O N S T R U C T I O N C O N S I D E R A T I O N S (2 0 3

being done as planned. If errors are detected during segment fabrication,

corrections can be built into subsequent segments as is done with cast-
in-place segmental construction.
If, during erection of the precast segments, it is found the anticipated
camber is not being achieved with the desired accuracy, the position of a
segment with respect to the previously erected one can be altered by
placing shims of metal or wire fabric in the joints in addition to the
epoxy adhesive. Alternatively, in some instances epoxy mortar has been
used to alter the position between precast segments. These procedures
have not frequently been used in precast segmental construction and
should not be needed if proper attention is directed to segment fabrica-
Another means of adjusting for deviations between the actual and an-
ticipated positions of completed cantilevers is shown in Fig. 6.23. The
procedure cannot be used if the superstructure and substructure are
permanently connected from the time the pier segment is first placed.
The procedure consists of providing space for hydraulic jacks, which
can be conventional hydraulic jacks or flat jacks, in such a manner that
the structure can be slightly raised, lowered or tilted and then shimmed
into the desired position before the permanent superstructure-
substructure connection is installed or completed.

6.6 Quantity of Prestressing Material

Prestressing steel, ducts and anchorages constitute important cost items

in prestressed concrete bridges. The number of anchorages required is
an especially important parameter for use in comparing the relative costs
of two or more designs. Each anchorage represents a cost made of the
1 . The cost of the anchorage itself.
2 . The formwork required to accommodate the anchorage.
3 . The labor required for placing the anchorage as well as for stress-
ing and grouting the tendon that terminates in the anchorage.
4 . Providing the necessary permanent protection for the anchorage
(concrete, formwork, curing, etc. for the concrete cover).
When the balanced cantilever erection technique is used, a consider-
ably greater number of anchorages is required than is needed when con-
ventional continuous cast-in-place construction methods are employed.
This is due to high cantilever dead load moments and the sequential
construction process that requires a few tendons to be stressed at each
segment. For the segmental construction technique to be the more
2 0 4 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

economical, the cost of the greater number of prestressing anchorages

must be offset by savings in falsework costs and savings which might
occur as a result of more rapid erection. A comparison of the quantities
of prestressing materials required for different solutions for a bridge
superstructure 600 feet long and 32.5 feet wide is shown in Table 6.1.




1 180-240-180 CIP Box 48 174,000

2 170-260- 170 CIP Segmental 216 156,000

3 90- 140- 140- 140-90 Segmental 232 120,000
*12+%” 4 2 7 0 strand anchorages.
- - -
A Long-Term
Concrete *

In the case of Portland Cement Concrete, class 325 or 400, hardened

under normal conditions, (in particular without heat curing), and subjected
to working stresses at most equal to 0.35 o’j**, in the absence of more
precise experimental results, the long-term deformations of the concrete
may be evaluated as follows:

1. Deformation due to shrinkage.

For concretes which have been protected against excessive loss of

moisture at early age, the shrinkage, as a function of time, may be deter-
mined using the expression l r r (t), in which r (t) is the function of time
defined below in section 1.5 and l r is the product of the four factors given in
the following paragraphs.
l r = kbeckekp

*From Reference 14, with minor editing.

**u’j = Compressive strength of concrete or the age of j days.

2 0 6 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

1.1 The coefficient kb depends on the composition of the concrete; it is

defined in Fig. A.I.

0.2 0.4 0.6 0.6

Fig. A.1 Chart showing relationship between coefficient kb and water-cement ratio for var-
ious cement contents. (Note that 100 Kg/m3 equals 169 pounds per cubic yards
or 1.79 sacks per cubic yard).

1.2 The coefficient cc depends on climatic conditions; it is defined in

Fig. A.2.

100 90 80 70 60 50 40
Relative air humidity, in %

Fig. A.2 Shrinkage coefficient EC shown as a function of relative humidity of average am-
bient air during service.

For heated floors, ovens etc., the values of l c should be drawn from

1.3 The coefficient kel depends on the theoretical thickness of the

member; it is defined in Fig. A.3.
k .I

Fig. A.3 Coefficient kel versus the theoretical thickness em.

The theoretical thickness em is defined as the quotient of the area B of the

section divided by the semi-perimeter p/2 in contact with the atmosphere;
for a section with one dimension very large with respect to the other, the
theoretical thickness is very close to the actual thickness.
If the dimensions are not constant along the member, it is possible to take
a theoretical thickness which is an average, paying particular attention to
sections subjected to the maximum stresses.

1.4 The coefftcient kp depends on the percentage of steel L = $, the

ratio of the sectional area of all the longitudinal steel (provided that it is
bonded) to the transverse sectional area of the member. It is expressed by
the formula:

kp= ’
1 + nf%
with n = 20, having regard to the effects of creep.
2 0 8 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

1.5 The law r(t), expressing the development of long-term deformation

with time, as a function of the time, is given in Fig. A.4.


0.8 i-- - - -
/ I -4 0.89




1 3 7 14 28 90 180 365 days -
1 2 5 years

Fig. A.4 Rate of deferred deformation (shrinkage and creep) with respect to time. Time is
actual time from application of load in the case of creep and is “theoretical time”
(defined in text) for shrinkage.

In order to take into account the size of the member, instead of the actual
time t (age of the concrete) the “theoretical time” is used in the diagram,
obtained from the following expression:

ti = t d--
in which the theoretical thickness em is expressed in cm.
That part of the deformation due to shrinkage in any interval of time
(t - ti) is equal to:
l r [r(t) - r(t)]

At an early age, the shrinkage of concrete which is protected is lower

than that in concrete which is not protected, (which is of importance in the
avoidance of cracks in concrete which is young and thus has a low
strength). This difference decreases and finally vanishes with time. The
phenomenon is less noticeable for very thick members.

2. Deformation due to creep.

As a first approximation, it may be assumed that creep is of a linear
nature, in evaluating the order of magnitude of long-term deformation due
to creep. This assumption, in the case of a constant stress m’b, leads to the
calculation of the final creeep deformation et-~ from the formula:
en = - Kn . r(t)
in which cib is the compressive stress in the concrete, Ei is the instan-
taneous elastic modulus of the concrete at the age of j days, r(t) is the
same function of time used in the shrinkage and Kn is the product of the
four coefficients described in the following paragraphs.
Kn = kbkckdkez
2.1 The coefficient kb depends on the composition of the concrete; its
value is identical with that given for shrinkage (see Section 1.1).
2.2 The coefftcient kc depends on the climatic conditions; it is defined
in Fig. A.5.
Air relative moisture



100 90 80 70 60 50 40

Fig. A.5 Creep ratio kc versus relative humidity of average ambient air during service.
2 1 0 )H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

2.3 The coefftcient kd depends on the hardening of the concrete at the

time of loading; it is defined in Fig. A.6, as a function of the age t of the
concrete at loading.

2 I

-1 3 7 14 28 90 180 360 days
Age at loading
A.6 Coefficient kd versus the age of the concrete at loading for ambient temperature of 20” C
(W F.). (See text for adjusting for different average ambient temperatures.)

The values in the diagram correspond to an average surrounding temper-

ature of 20°C (68°F); where the temperatures are below this value, a
correction should be made to the age of loading and the new value tc is given
in the following formula, valid for temperatures lying between -10 “C
(14°F) and + 20 “C (68°F);
tc = ZAt(e + 10)
in which:
tc represents the corrected loading age expressed in days,
At represents the number of days during which the hardening has oc-
curred at a temperature 8 expressed in degrees centigrade.

2.4 The coefficient kez depends on the theoretical thickness em of the

member defined in Section 1.3; it is given in Fig. A.7.





Fig. A.7 Theoretical thickness coefficient for creep, kez, versus theoretical thickness.

2.5 The law r(t) is the same as for shrinkage, but the number of days is
counted from the application of the load; on the other hand, the actual time*
is taken into account, since the development of creep with time is less
sensitive than that of shrinkage, to the influence of a reduction in the
theoretical thickness of the member.
2.6 At a given time t counted from the application of loads, the effect of
a stress dbj applied at a momentj and subjected at any instant i to variations
in intensity AC'bi (algebraically) may be taken as equal to:

et-~ = kbkcke &$ . kdjr(t - j) + zq kdir(t - i) ]


an expression in which the moduli Ej and Ei are both instantaneous moduli,

the subscripts i and j referring to the moments designated above.
*In contrast to the theoretical time.
B Standard
Shapes of


/- -


Fig. B.l ASSHTO-PCI Standard Bridge Beams types I through VI.


Moment of
Beam Type Area Inertia Yt Span
in2 in4 .i n feet

I 276 22,750 15.41 30- 45

II 369 50,980 20.17 40- 60
III 560 125,390 24.73 55- 80
IV 789 260,730 29.27 70-100
V 1013 521,180 31.04 90-120
VI 1085 733,320 35.62 110-140


1 l’-7” 1
Fig. 8.2 California standard l-girders.


Moment of
Depth Area inertia Yf Span
in in2 id in Feet

36 432 63,300 18.9 50- 55

42 474 95,000 22.0 55- 65
48 516 137,300 25.2 65- 75
54 558 187,800 28.3 75- 80
60 604 248,600 31.4 80- 90
66 642 318,000 34.4 90-100
72 684 400,600 37.5 100-I 10

Fig. 8.3 Colorado standard girders.



Moment of Maximum
Depth Area Inertia Yt Span
in in* in4 in Feet

54 630.5 242;59 1 21.33 105

68 647.5 413,725 34.08 125
72 759.5 538,144 32.76 134

24 ,

TYPE 40 TYPE 60 TYPE 80 TYPE 100 TYPE 120

Fig. 8.4 Standard precast concrete girders, State of Washington.


Moment of Nominal
Depth Area Inertia Yt Span
in in2 id in Feet

32.0 253 31,000 16.84 40

42.0 332 70,100 23.31 60
50.0 476 154,900 27.47 80
58.0 546 249,000 30.10 100
73.5 626 456,000 37.90 120
C Analysis of Statically
with the Method of
Support Constants*

Definition of the Constants

1. Elasticity of the joints with respect to the beam (K).
Consider the structural system shown in Fig. C. 1 which incorporates a
beam spanning between joints 1 and 2. Rotations 81 and 02 occur at joints 1

Fig. Cl Beam spanning between two joints.

*Private communication from Gerard Sauvageot, Project Engineer, Europe Etudes, Paris,
2 1 8 )H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

Fig. C.2 Freebody diagram for support members at joint 1.

and 2 respectively, due to the applications of loads to beam 12. The

rotations 81 and 6% induce moments Ml2 and M21 in beam 12 atjoints 1 and 2
respectively. Moments of equal magnitude, but opposite direction are
introduced in the supporting members at joints 1 and 2. The elasticity of
joint 1 with respect to beam 12 is defined as follows:

K l/12 = + (1)

The moment Ml2 is applied to the joint 1 as shown in Fig. C.2. In a similar
manner the elasticity of joint 2 with respect to beam 21 is as follows:
K U2I = -

Fig. C.3 Freebody diagram for joint 1.


The stiffness (R) is reciprocal of the elasticity. The stiffness ofthe joint 1
with respect to beam 12 is equal to:
l/l2 = -

I and that for joint 2 with respect to beam 21 is:

It should be recognized that stiffnesses, but not elasticities, can be added.

Considering the freebody ofthe joint 1 shown in Fig. C.3, the member 13
has a moment Ml3 at joint 1 equal to:

Ml3 = 4Rl3 (5)

In a similar manner for member 14, there is a moment at joint 1 equal to:

M,, = 4R,, (6)

and for member 15 the moment is:

MH = &Ra (7)

Finally, one can relate the sum of the moments in the members which
constitute the supports at joint 1, to the moment MI-Z, as follows:

Mlz = M13 + Mlr + MI, (8)

From which comes the equivalent expression:

&k/,2 = 4 (R,3 +R,4 +Ris) (9)

R l/l2 = R13 + RI, + R,, (10)

Hence, the stiffness of the joint 1 with respect to the beam 12 is equal to the
sum of the stiffnesses of all the members at joint 1 except for beam 12.
Knowing the stiffnesses of the members which frame into the joints, one
can determine the joint stiffness and elasticity.
2 2 0 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

2. Elasticity of the beam with respect to the joints (k).

If one assumes beam 12 of Fig. C. 1 is hinged at joint 1 and a moment is
applied at end 1 as shown in Fig. C.4, the elasticity of the beam with respect
to the support is:
k IUI = -
and the stiffness of the beam with respect to the support is:
rlul = -
In a similar manner the elasticity and stiffness of the beam with respect to
the joint 2 are:
k21/z = - (13)

r2i/2 = -
e2 (14)

3. Properties of the Beam

If beam 12 is simply supported and a unit moment is applied at end 1,
rotations which are designated a and bare obtained at ends 1 and 2 as shown
in Fig. C.5. It can be shown that:

a=-1 ’ (l-x)2 dx
fi (13
I2 s0

Fig. C.4 Freebody diagram with hinge at joint 1.


Fig. C.5 Freebody diagram for beam 12 simply supported with a unit moment at 1.

where x is measured from 1 toward 2; and

dx (16)
N-x) E

In which E and I are the elastic modulus and the moment of inertia of the
beam respectively.
In a similar manner if a unit moment is applied at end 2 as shown in Fig.
C.6, rotations equal to b and c are obtained at ends 1 and 2 respectively.
The value of c is:
1 ’ xz dx
c =
F s0 fi (171

For a member of constant moment of inertia:

Fig. C.6 Freebody diagram for beam 12 simply supported with a unit moment at 2.
2 2 2 (H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

The rotation at the ends opposite to that at which the unit moment is
applied is equal to b in each case. This can be explained by Maxwell’s Law
of Recriprocal Deflections.
For loads applied to the simply supported beam 12, rotations occur at
ends 1 and 2 which are designated WI and ~2 respectively. This is illustrated
in Fig. C.7.

Fig. C.7 Simply supported beam with transverse loads.

4. End moments on beam 12 including the effects of continuity.

The rotations at the ends of beam 12, including the effects of loading
within the span as well as the end moments due to continuity, as shown in
Fig. C.8, can be expressed as follows:

8, = 01 + aMl2 - bM2r (19) ’


e* = 02 - bM12 + cMzr (20)

substituting equations (1) and (2) for & and 02 respectively in equations (19)
and (20) and solving for Ml2 and MU, one obtains:
( c + Ku24 01 +boz
M12 = (a + KIN) (c +Ku21) -b2

bo, + (a + K/12) wz
M21 = - (a + KI,I~) (c +Ku21) - b2 (22)

In the equations (21) and (22) “beam” moment sign convention, as shown
in Fig. C.9 is used.

Fig. C.8 Rotations at the ends of beam 12 resulting from transverse loads and end mo-

o- W+
c - :’ i: M-uM->

Fig. C.9 Sign convention for rotation and moment (beam convention).
2 2 4 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

5. Computation of the elasticities of the beam.

With the elasticity of the joint 1 (KIN) being known, the elasticity of the
beam at the support 2 (kzvz) is determined by applying amoment mz in beam
21, atjoint 2 (hinge assumed at 2 in beam 21). This is illustrated in Fig. C. 10.
The moment m2 induces a moment of Ml2 in beam 12 at joint 1. (See Fig.

01 = Km2 M12 (23)

8, = kzl/2 m2


8, = -aMI - bm, (25)

B2 = bM,2 + cm2 (26)

from which

KIIIZM~~ = - aMl2 - bm2 (27)


h/2 m2 = bM12 + cm2 (28)

Fig. C.10 Frame with hinge assumed at 2.



(a + KdM12 = -bm2 (29)


(c - km) mz = - bM12 (30)

which combined gives

(a + Kw)(c - kzl/z) - b2 = 0 (31)

From which the general equations for the elasticities of the beam are
determined to be:
k 21/z = c - (32)
a + KVIZ
k IUI =a- -
c + Kz21 (33)

6. Carryover Factors.

Rearranging equation (29), one obtains:

Ml2 = -b em2 (34)

a + KUIZ
Therefore, the carryover factor from 2 to 1 is:

c,, = b (35)
a + h12


Fig. C.11 Relationship between joint moment mz at beam moment MIZ.
2 2 6 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

and from 1 to 2
Cl, = b
c + Ku21
The carryover factors are illustrated in Fig. C.12.
The equations for the end moments in the continuous beam can be
written from equations (21) and (22) as:
M,* = + cT . (;)I-+cc12cu2 (37)
12 21

M,, = _ c<2 . T’i”;: ‘c”’

12 21

Elasticity of the Joints: The elasticity of the joint is computed as the
reciprocal of the sum ofthe stiffnesses of the members which frame into the
joint, neglecting the beam under consideration. The elasticity of a member
is equal to the angle of rotation of the member under a unit moment,
assuming the member is hinged at the joint under consideration. For

C 21

Fig. C.12 Diagram showing definition of carry-over factors.


prismatic members, the relative stiffnesses for members hinged and fixed at
their far ends are:

Hinged members R = F (39)

Fixed members R = F (40)

Elasticities of the Beams

k I?/1 =a- b2 (41)

c + Ku21

k 21/z b2
= c -
a + KI/I~

Carryover Factors

c,, = b (43)
c + Kv21

cpl =a + bKlj12
End Moments

Ml2 = + k . 01 + Cl202
b 1- Cl2CPl


Example Problem
Compute the bending moments due to the applied live load of 10 klf as
well as the secondary bending moments due to the prestressing force for
the structure shown in Fig. C. 13. The structure can be idealized as shown
in Fig. C. 14 where it will be seen the moments of inertia for the beams and
piers are 400 ft.4 and 100 ft.4 respectively.
The properties of the structure are determined by first finding the unit
moment coefftcients a, b and c using Eq. 18. These are as follows:
2 2 8 )H I G H W A Y B R I D G E S U P E R S T R U C T U R E S


70 *



Fig. C.13 Frame used in example calculation.

For piers 25 and 36 (taking E = 1)

a=c=2b= 2o = 0.0666
3 x 100

For beams 12 and 34 (taking E = 1)

a=c=2b= loo = 0.0833
For beam 23 (with E = 1)

a=c=2b= 130 = 0.108

Considering pier 25 (Fig. C.15, because joint 5 is hinged, (I&s2 = a~),
from Eq. 33:
ksu2 = a = 0.0666

In the case of pier 36 (Fig. C. 16), the joint 6 is fixed (K&x = 0) and from Eq.
km = 0.0666 - y=& = 0.05

I = 400 ft.* r/
1 2 3 45
7///f I=100 ft.’
/\5 6 /

Fig. C.14 Schematic layout for frame of Fig. C.13.


Fig. C.15 Freebody diagram for pier 25.

For beam 12, joint 1 is hinged (KI/IZ = m) and from Eq. 32:

kz,,z = 0.0833 - 0.04162 = 0.0833

0.0833 + m
For joint 2 with respect to span 23, by Eq. 10:

1 1
i&3 = iii,* + iii2
1 =- 1 - 1
i-G*3 0.0833 + 0.0666
Kux = 0.037

-. .-

Fig. C.16 Freebody diagram for pier 36.

2 3 0 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

The elasticity of span 23 with respect to joint 3, from Eq. 32 is:

kjyj = 0.108 = 0.0879
0.037 + 0.108

and the elasticity ofjoint 3 wit,h respect to span 34 is computed using Eq. 10
as follows:
1 1 1
E/34 =m+ 0.05

K3,34 = 0.032

Joint 4 is fixed. Therefore K4132 = 0. The elasticity of span 43 with respect to

joint 3 is:
k43,3 = 0.0833 - 0.0416* = 0.0625
0.0833 + 0

and the elasticity ofjoint 3 with respect to span 32, is;

1 1 1
G = -
0.0625 ‘+ 0.05

K3,32 = 0.0277
k3uz = 0.108 - 0.0542 = 0.0866
0.108 + 0.0277

from which

1 l = 0.0866
1 + 0.0666

Kz,z, = 0.0377

The joint elasticities can be summarized as follows:

K l/l? = m Kz/z, = 0.0377

KUZ3 = 0.037 K3132 = 0.0277
K3134 = 0.032 K4/43 = 0

1 -7
/ 55.6%

Fig. C.17 Freebody diagram for joint 2 left.

The distribution factors for each joint are determined from the stiffness of
the members. For moment M23 as shown in Fig. C.17:

kz,,z = 0.0833 h/2 = 12 dt, = 0.444

kzs,z = 0.0666 rz12 = 15 dzs = 0.556

Cr = 27
and for MZI as shown in Fig. C.18.

k23/2 = 0.0866 I-2312 = 11.54 dz3 = 0.435

kz/z = 0.0666 r25/2 = 26.54

15.00 deJ = 0.565

and for M32 as shown in Fig. C.19

k,.,,3 = 0.0625 r34/3 = 16 d34 = 0.444

kw3 = 0.050 rW3 = 20 d36 = 0.556

Zr = 36

and for M34 as shown in Fig. C.20

km = 0.0879 r313 = 11.38 d32 = 0.362

kw, = 0.050 r36/3 = 20 d36 = 0.638

Zr = 31.38
2 3 2 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

M \/

c 2 3 45
56.5% /
\ I\5
Fig. C.18 Freebcdy diagram for joint 2 right.

Fig. C.19 Freebody diagram for joint 3 right.

1 2 3

Fig. C.20 Freebody diagram for joint 3 left.

The carry over factors from right to left are computed from Eq. 35 while
that from left to right from Eq. 36. These are:

CZl = O-0416 = 0 Cl0 = 0.0416 = 0.344

0 . 0 8 3 3 + 00 0.0833 + 0.0377

c32 = o*w16 = 0 - 372 C 23 = 0.054 = 0.40

0.108 + 0.037 0.108 + 0.0277

c43 = 0.0416 = 0.36 c3, = 0.0416 = 0.50

0.0833 + 0.032 0.0833 + 0

This completes the computations of the properties of the structure.

Under the loading of 10 klf on span 23, the simple span values of Eoz
and EW can be shown to be:

EON = - 2288

EWQ = + 2288

in which a counterclockwise rotation is positive. Using Eq. 37 and 38, the

end moments in span 23 are found to be as follows:

- 2288 + 0.4 x 2288 = _ 1 1 9 1 13 K - Ft .

M23 = + s *
1 - 0.372 x 0.4

M32= - -. -
0.054 0.4 2288
1 - x0.372
0.372x +0.42288 = _ 12 , 50, K - Ft .

using the distribution factors computed above for M23 and h&2 one
obtains a distribution of moments as shown in Fig. C.21.
The secondary moments for the portion of the tendon in span 23 is
computed next. The same procedure should be used for the portions of the
tendon in spans 12 and 34 and the results of the three analyses added
together for the combined effect. Refer to Fig. C.22. Assuming the tendon

Fig. C.21 Moment diagram for frame with loading shown in Fig. C.13.
2 3 4 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S


Fig. C.22 Freebody diagram for prestressing between joints 2 and 3.

is on a trajectory of compounded second degree parabolas, the areas of the

moment curves are computed as follows (E = 1):
Between 12 and 34:

-= 5000x4 = -50
I 400

Area = - 50 x 30 x f = - 1000

Between 23:

-= 5ooox5 ~625
I 400 .

Area = 62.5 x 70 x t = + 2917

w2 = + 2917 - 1000 - 1000 = + 458


03 = - 458

and the moments M23 and M32 are as follows:

M23 = %&. 458 - 0.4 x 458 = 2224
1 - 0.372 x 0.4

M32 = - -0.4 . 0.372 x 458 - 458 = 2503

0.054 1 - 0.372 x 0.4

Fig. C.23 Secondary moments due to prestressing in span 23.

and the secondary moment diagram as shown in Fig. C.23 is found using
the distribution factors for moments M23 and M32.

Note: When the tops ofthe piers are free to translate horizontally, ageneral
method must be used for the computation of the piers. With reference to
Fig. C.24, it will be seen that the pier is subject to a bending moment of M
and a horizontal force Q which are applied at the top. The rotation at the
top due to M and Q is 8 while the displacement is equal to 6. Using the
A = the rotation due to the moment (elasticity for rotation.)
B = the rotation due to the horizontal force, which because of the
Maxwell Theory is equal to the displacement due to the moment.
C = the displacement due to the horizontal force (elasticity for dis-
E = Young’s modulus.

Fig. C.24 Freebody diagram for typical pier.

2 3 6 (H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

it can be shown that

E8 = AM + BQ (47)

E6=BM+CQ (48)
in which
h dx
A = (49)
s3 \

B= h xdx
s0 I


If the top of the pier cannot move, E6 = 0, Q = - $ M (52)

EB=M(A-!!&kM (53)

k=A-E (54)

For a constant moment of inertia,

A=+ (55)

fj=hZ (56)


In the case of pier 36, A = 0.20, B = 2.00 and C = 26.67 and

k = 0.20 - & = 0.05

This is the same value computed above.


For the case where the top of the pier cannot deflect horizontally
(ES=O) it can be shown that the distance x from the top of the pier to
the point of zero moment is:
x =- (58)
Under the effect of a lateral load, the tops of the piers deflect as
shown in Fig. C.25. Considering the joint at the top of the pier on the
left side, it can be shown that for the pier Eq. 47 applies while the rota-
tions result in the following relationships for the beams:

span 1: E8 = - GM (59)

Span 2: E8 = - (a2 - bz) M” (6’3

and for the joint:

M = M’ + M” (61)

Fig. C.25 Deflected shape of a frame under lateral load.

From the above one obtains:

M, = (* - b2) M (62)
cl + (a2 - bz)

M” = Cl M
cl + (az - b2)

A= l +’ (64)
(al-b2) cl
2 3 8 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

one can show that

M = -BAQ (65)
Ah + 1
and the distance x from the top of the pier to the point of zero moment

x =Bh (66)
AX + 1
The effect of a length change, which may be due to a combination of
creep, shrinkage or temperature-induced strain, the piers deflect as
shown in Fig. C.26. Eq. 47 remains correct for the pier and the rota-
tional relationships for the beams at the top of the left pier are as fol-

Fig. C.26 Deflected shape of a frame due to change of beam length.

Span 1: E8 = c,M’ (67)

Span 2: E8 = -(an + b2) M” (68)

and for the joint Eq. 61 applies. From this one can show that
M, = (a2 + W M
cl + (ao+ b2)
M” = Cl M
cl + (a2 + b2)
and if
*’ = (a2 : b2) + z (71)

M = -BA’Q (72)
Ah’ + 1

and the distance x from the top of the pier to the point of zero moment

x= BX’
AA’ + 1
Finally, it can be shown that the relationship between the shear force Q
and the deflection at the top of the pier is:

Q = ESB,~, (74)
’ - AA’ + 1
D Thermal
Stresses in
Concrete Bridge

When the temperature distribution within a member is known, the ther-

mal stresses in the section can be easily computed (Ref. 23). Priestley has
shown if plane sections are assumed to remain plane, for a section as shown
in Fig. D. 1 subject to the temperature distribution shown in Fig. D.2 the
thermally induced strains and stresses are as shown in Fig. D.3. The
equations for stress and for force and moment equilibrium are as follows:

fy = E (~1 + ~2 f - CY ty)

ty by dy

in which the terms are defined in Figs. D. 1 through D.3 and (Y is the linear
coefficient of thermal expansion, A is the area of the section and I is the
moment of inertia of the section about the horizontal axis pasing through
the centroid. The procedure consists of solving equation (D.3) which gives
the curvature of the section and the value of ~2. With ~2 known,equation
(D.2) can be solved for the value of El and the stresses found from equation
(D.l). The curvature found with equation (D.3) is used to determine the
additional stresses induced by continuity in structures which are continu-
ous. Curvature is equal to M/E1 and deflections can be computed there-
The bridge designer generally does not have precise data relative to the
extreme temperature gradient a particular bridge will experience in ser-
vice. For this reason an approximate method of analysis will yield results
that are adequate for most design work. The approximate method consists


Fig. D.l Cross section of a tubular beam.


Fig. D.2 Temperature gradient in beam of Fig. D.l.


(a) . W
Fig. D.3 Strain and stress distributions in beam of Fig. D. 1 due to temperature gradient of Fig.
2 4 4 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

~ Fig. D.4 Typical bridge cross section.

of assuming the top deck is raised to a uniform temperature higher than the
other parts of the cross section. If unrestrained, the top deck would
experience an increase in length as a result of the temperature increase.
The expansion of the top deck is resisted by the webs and bottom flange.
The first step in computations with the approximate method consists of
computing the forces that would exist in the top flange due to the increase in
temperature if the flange were completely restrained. The force results in a
compressive stress in the top flange. The effect ofthe force in the top flange
on the section as a whole is computed by applying a tensile force equal in
magnitude to the compressive force, applied at the centroid of the top
flange. The sum of the stresses from the two forces gives the approximate
stresses due to thermal effects in the section.

(a) W w

Fig. D.5. Stresses in the bridge of Fig. D.4, by approximate calculation, for a temperature
differential of 30°F.

A A i
210 210 t, 155 w

Fig. D.6 (a) Elevation of beam continuous over four spans and (b) secondary moments in the
beam due to differential temperature of 30°F.

As an example, consider the bridge cross section shown in Fig. D.4.

Assume the top flange temperature is 30” F. higher than that of the webs
and bottom flange. The compressive stress in the top flange, if fully
restrained, would be:

fc = (At) (a) (EC) (D.4)

in which (Y is the linear coefficient of thermal expansion and EC is the elastic

modulus of the concrete. If a = 0.0000065 and EC = 3,000,OOO psi, fc = 585
psi. The force in the top flange would be:

p = 585 x 5000 = 2925 kips

The tensile force applied 5.29 inches from the top of the section produces
the stresses shown in Fig. D.5(b) and the combined stresses are as shown
in Fig. D.~(c). These are the approximate stresses that would exist if the
beam were a single simply-supported span.
If made continuous over four spans as shown in Fig. D.6(a), it can be
shown that secondary moments and reactions as shown in Fig. D.5(b)
would exist. The secondary reactions are those required to prevent the
beam from deflecting upward from its supports as a result of the thermal
gradient. The secondary moment at the first interior support results in a
stress of 314 psi (compression) in the top fiber and 510 psi (tension) in the
bottom fiber. The net stress in the member due to the effect of the differen-
tial temperature would be those obtained by combining the stresses shown
in Fig. D.5(c) with those resulting from the secondary moment of Fig. D.6.

1. Ferguson, P. M. Reinforced Concrete Fundamentals. New York: John Wiley

& Sons, 1959.
2. Libby, J. R. Modern Prestressed Concrete Design Principles and Construc-
tion Methods. New York: Van Nostrand Reinhold Co., 1971.
3. “Standard Specifications for Deformed and Plain Billet-Steel Bars for Con-
crete Reinforcement.” ASTM Designation A61575. American Society for
Testing and Materials, Philadelphia, 1975.
4. “Standard Specifications for Uncoated Seven-Wire Stress-relieved Strand for
Prestressed Concrete.” ASTM Designation A41674. American Society for
Testing and Materials, Philadelphia, 1975.
5. “Standard Specifications for Uncoated Stress-relieved Wire for Prestressed
Concrete.” ASTM Designation A421-74. American Society for Testing and
Materials, Philadelphia, 1975.
6. American Association of State Highway and Transportation Officials. Stan-
dardSpecificationsforHighway Bridges (and 1974and 1975 Interim Specifica-
tions, Bridges), 11th ed. Washington, D. C., 1973.
7. AC1 Committee 318. Building Code Requirements for Reinforced Concrete
(AC1 318-71). American Concrete Institute, Detroit, 197 1.
8. AC1 Committee 443. “Preliminary Design and Proportioning of Concrete
Bridge Structures.” ACI Journal, Proceedings V. 70, No. 5, May 1973, pp.
9. AC1 Committee 443. “Analysis and Design of Reinforced Concrete Bridge
Structures.“AClJournal, Proceedings V. 71, No. 4, April 1974, pp. 171-200.
2 4 8 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

10. Seni, Alfio. “Comparison of Live Loads Used in Highway Bridge Design in
North America with Those Used in Western Europe.” Second International
Symposium on Bridge Design, Volume 1. Detroit: American Concrete Insti-
tute, 1971, pp. l-34.
11. Rajagopalan, K. S. “Comparison of Loads Around the World for Design of
Highway Bridges.“Second InternationalSymposium on Bridge Design, Vol-
ume 1. Detroit: American Concrete Institute, 1971, pp. 35-48.
12. Libby, pp. 40-58.
13. Branson, D. E. and Kripanarayanan, K. M. “Time Dependent Concrete
Properties, Prestress Loss, Camber and Deflection of Noncomposite and
Composite Structures of Different Weight Concretes.” Sixth Congress, Fed-
eration lnternationale de la Precontrainte, Prague, June 1970.
14. Association Francaise du Btton. Conception et Calcul du Beton Prtcontraint
(Conception and Design of Prestressed Concrete), Instruction provisoire du 13
aout 1973, Appendix 1, Paris, France, (in French).
15. Scordelis, A. C. and Davis, R. E. “Stresses in Continuous Concrete Box
Girder Bridges.” Second International Symposium on Bridge Design, Volume
1. Detroit: American Concrete Institute, 1971, p. 284.
16. Westergaard, H. M. “Computation of Stresses in Bridge Slabs Due to Wheel
Loads.” Public Roads, March 1930.
17. Pucher, Adolf. Einflussfelder Elastischer Platten (Influence Surfaces of Elas-
tic Plates). New York: Springer-Verlag, 1964 (in German and English).
18. Homberg, Hellmut. Fahrbahnplatten mit veraenderlicher Dicke (Decks with
Variable Thickness), Volumes I and II. New York: Springer-Verlag, 1968, (in
19. AC1 Committee 318, p. 23.
20. American Association of State Highway and Transportation Officials, p. 33.
21. Association Francaise du Btton, p. 189.
22. Ibid, pp. 13-16.
23. Priestley, M. J. N. “Model Study of a Prestressed Concrete Box-Girder
Bridge Under Thermal Loading.” Proceedings ofthe Ninth Congress ofthe
IABSE, Amsterdam, May 1972.
24. Radolli, M. and Green, R. “Thermal Stresses in Concrete Bridge
Superstructures-Summer Conditions.” Presented at the 54th Annual Meet-
ing, Transportation Research Board, Washington, D.C., January 1975.
25. Bouwkamp, J. G., Scordelis, A. C. and Wasti, S. T. Structural Behavior ofa
Two Span Reinforced Concrete Box Girder Bridge Model, Volume I, UC-
SESM Report No. 71-5. Berkeley: College of Engineering, Office of Research
Services, University of California, April 1971.
26. Scordelis, A. C., Bouwkamp, J. G. and Wasti, S. T. Structural Behavior ofa
Two Span Reinforced Concrete Box Girder Bridge Model, Volumes II and III,
UC-SESM Report No. 71-16 and 71-17. Berkeley: College of Engineering,
Office of Research Services, University of California, October 1971.
27. Kristek, V. “Theory and Research on Thin-Walled Prestressed Concrete
Beams.” Proceedings of the Sixth Congress, Federation International de la
Precontrainte, Prague 1970, pp. 164-173.
28. Nayak, G. C. And Davies,J. D. “Influence Characteristics for Slab Bridges.”
Second International Symposium on Bridge Design, Volume I. Detroit:
American Concrete Institute, 1971, pp. 75-l 16.
R E F E R E N C E S 12 4 9

29. Douglas, M. R., Parekh, C. J., and Zienkiewicz, 0. C. “Finite Element

Programs for Slab Bridge Design.” Second International Symposium on
Bridge Design, Volume 1. Detroit: American Concrete Institute, 1971, pp.
30. Motajemi, D. and VanHorn, D. A. “Theoretical Analysis of Load Distribu-
tion in Beam-Slab Bridges.” Fritz Engineering Laboratory Report No. 315.9.
Bethlehem, Pennsylvania: Fritz Engineering Laboratory, Lehigh University,
31. Courbon, J. “Calcul des Ponts a Poutres Multiples Solidarisees par des En-
tretoises,” (Design of Multi-Beam Bridges with Intermediate Diaphragms).
Annales des Ponts et Chausees, Paris, Nov./Dee. 1940, (in French).
32. AC1 Committee 443, pp. 171-200.
33. Libby, pp. 99 and 193.
34. Burdette, E. G. and Goodpasture, D. W. Final report on “Full-Scale Bridge
Testing.” Tennessee Department of Transportation and Federal Highway
Administration, 1971.
35. Burdette, E. G. and Goodpasture, D. W. “Test to Failure of a Prestressed
Concrete Bridge.” PC1 Journal, V. 19, No. 3, May-June 1974, p. 92.
36. Libby, J. R. Discussion of the paper “Test to Failure of a Prestressed Con-
crete Bridge,” by Burdette, E. G. andGoodpasture, D. W. PCZJournaf V. 20,
No. 1, Jan.-Feb. 1975, p. 101.
37. Anderson, Arthur R., private communication, January 2, 1975.
38. AC1 Committee 443. “Analysis and Design of Reinforced Concrete Bridge
Structures-Prestressed Concrete.” To be published in the ACI Journal.
39. Timoshenko, S. “Strength of Materials,” Part II, 3rd ed. Princeton, New
Jersey: D. Van Nostrand Company, Inc., 1956, pp. 247-254.
40. Oden, J. T. “Mechanics of Elastic Structures.” New York: McGraw-Hill
Book Company, 1957, pp. 53-57.
41. Timoshenko, S., p. 68.
42. Brown, Robert C., Jr., Bums, Ned H. “Computer Analysis of Segmentally
Erected Bridges.” Journal of the Structural Division, Proceedings of the
AmericanSocietyofCivilEngineers, V. 101, No. ST4, April 1975, pp. 761-778.
43. Oden, J. T., pp. 118-122.
44. Western Concrete Reinforcing Steel Institute. “Post-Tensioned Box Girder
Bridges.” Burlingame, California, 1969.
45. Libby, J. R. Modern Prestressed Concrete Design Principles and Construc-
tion Methods. New York: Van Nostrand Reinhold Co., 1971, pp. 480487.
46. Muller, J. “Long-Span Precast Prestressed Concrete Bridges Built in Can-
tilever.” First International Symposium on Concrete Bridge Design, AC1
Special Publication SP-23, Detroit, 1969, pp. 705-740.
47. Subcommittee 5, AC1 Committee 435. “Deflection of Prestressed Concrete
Members.” ACZ Journal, Proceedings V. 60, No. 12, December 1963, pp.
48. Thenoz, M. “Redistribution des Efforts par Fluage dans les Ponts Construits
par Encorbellement,” (Redistribution of Moments due to Creep in Bridges
Constructed in Cantilever). Contributions Techniques Francaises, 7~ Con-
gress de la Federation Internationale de la Precontrainte, New York, 1974,
pp. 25-27, (in French).
49. Muller, J. “Concrete Bridges Built in Cantilever.” 15lst Ordinary General
Meeting, Societe des Ingenieurs Civils de France, British Section, London,
May 24, 1963, pp. 22-27.
250 ) H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

50. Association Francaise du BCton, pp. 82-84.

51. AC1 Committee 318, pp. 34-38.
52. Tang, Man-Chung. “Shear Design of Large Concrete Box Girders.” Shear in
Reinforced Concrete, V. 1. Detroit: American Concrete Institute, 1974, pp.
53. ACI-ASCE Committee 426. “The Shear Strength of Reinforced Concrete
Members.” ACT Manual of Concrete Practice, Part 2, 1974, pp. 426-21.
54. American Association of State Highway and Transportation Officials. Interim
Specifications Bridges, Washington, D. C., 1975, p. 80.
55. Bassi, K. G., Lin, W. L. and Richardson, B. S. “Continuous Post-Tensioned
Torsionally Stiff Concrete Bridges.” First International Symposium, Con-
crete Bridge Design, AC1 Special Publication SP-23, Detroit, 1%9, pp. 563-
56. Witecki, A. A. “Simplified Method for the Analysis of Torsional Moment as
an Effect of a Horizontally Curved Multi-Span Continuous Bridge.” First
International Symposium, Concrete Bridge Design, AC1 Special Publication
SP-23, Detroit, 1969, pp. 193-204.
57. Witecki, A. A., p. 198.
58. Bassi, K. G., Lin, W. L., and Richardson, B. S., p. 570.
59. American Association of State Highway and Transportation Officials. Stan-
dard SpeciJications for Highway Bridges, 1 lth ed. Washington, D. C., 1973,
pp. 238-240.
60. Mathivat, J. “Structures de Piles AdaptCes a la construction par Encorbel-
lement,” (Piers Adaptable for Use in Cantilever Construction). Contributions
Techniques Francaises, Seventh Congress of the International Federation of
Prestressing, New York, 1974, pp. 36-53, (in French).
61. Muller, J., pp. 20-22.
62. American Associationof State Highway and Transportation Officials. Znterim
Spectjkations Bridges, Washington, D. C., 1975, pp. 6-13.
63. American Association of State Highway and Transportation Officials. Interim
Specifications Bridges, Washington, D. C., 1975, pp. 14-29.
64. Imbsen, Roy, and Gates, James. “Recent Innovations in Seismic Design and
Analysis Techniques for Bridge Structures.” Proceedings, 42nd Annual Con-
vention, Structural Engineers Association of California, Coronado, Califor-
nia, October 1973, pp. 81-135.
65. Department of Transportation, Business and Transportation Agency. Stan-
dard Specifications, State of California, January 1975, pp. 249-251.
66. Muller, J. “Ten Years Experience in Precast Segmental Construction.” Pro-
ceedings, 42nd Annual Convention, Structural Engineers Association of
California, Coronado, California, October 1973, pp. 16-38.
67. Muller, J. “Precast Segmental Bridges-Conception and Design Fundamen-
tals.” Convention of the Prestressed Concrete Institute, Chicago, 1973.

A Effective deck span, 9

AASHTO, 2 Shear design, 145
AASHTO Specification, 2, Committee 435, Deflection, 98
decks, 7, 46 Committee 443, Bridge Design, 3
diaphragms, 44, 76 Allowable stresses, 13
elastomeric bearing pads, 154, 157
effective deck span, 9 B
emperical coefficients, 5, 65, 66 B-3 Viaduct, 19
intermediate diaphragms, 76 Box-girder bridge, 15, 18, 57
lane width, 5 flexural analysis, 59, 66
live load, 3, 94, 117 Bridge, types, 2, 15
load distribution, 37, 39, 40 Bridge, wide, 33
minimum slab thickness, 73
multi-beam bridge, 25
segmental bridges, 18 EALTRANS, 79, 187
seismic forces, 173 Cam&r
shear design, 145, 149 box-girder bridges, 197
spread box-beam bridge, 26 girder bridges, 5 1
tensile stresses, 13 segmental box-girder bridges, 197
AASHTO-PC1 beams, 43, 213 Cantilever
AC1 erection, 85
Committee 209, creep and shrinkage, 197 moments, 88
Committee 3 18 tendons, 141
Building Code Requirements, 2 Chillon Viaduct, 171

2 5 2 1H I G H W A Y B R I D G E S U P E R S T R U C T U R E S

Choisy-Le-Roi Bridge, 170, 189 Emperical coefficients

Concrete, 197, 209 deficiencies in, 7
creep, 4, 209 deck design, 5
fmishes, 187 box-girder bridges, 38
shrinkage, 4, 197, 205 girder bridges, 37
Continuity multi-beam bridges, 26
box-girder bridges, 57, 73 relationships, 25
girder bridges, 48 segmental box-girder bridges, 18
tendons, 88, 141 spread box-beam, 26, 38
segmental box-girder bridges, 85, 193 Erection
Construction considerations precast girders, 188
box-girder bridges, 73 precast segments, 188
general, 175 sequence, 88, 192
girder bridges, 50
segmental box-girder bridges, 141 F
Courbon, P., 34, 40 Falsework, 176
Creep redistribution of moment, 85, 96 Forms
Curved bridges, 150 cast-in-place segments, 181
general, 179
precast segments, 182
D French Code
Decks classes of structures, 13
box-girder bridges, 60, 67 creep, 4, 205
design, 7 deck span, 13
elastic analysis, 9
shrinkage, 4, 205
emperical analysis, 7
girder bridges, 46 G
post-tensioned, 102 Gantries, 85, 190
segmental box-girder bridges, 84 Girder
span, 9, 13 design, 42
variable thickness, 9 dimensions, 43, 213
Design proportioning, 42
criteria, 2 Girder bridges, 15, 29
methods, 5 design, 42
loads, 3
Diaphragms H
box-girder bridges, 59, 76 Harbor Drive Overcrossing, 53
Courbon design method, 34, 40 Hinge, 84, 134
end, 17, 23, 27, 76, 113 Homberg, H., 9, 25, 46, 100
general, 29, 52 Horizontal curvature, 150
girder bridges, 16, 27, 34, 44, 52
intermediate, 17, 23, 27, 34, 44, 76, 84,
96, 105 I
post-tensioned, 45 Influence lines, 33
segmental box-girder bridges, 84, 96, 105 Influence surfaces, 9, 100
skew girder bridges, 52
structural steel, 46 L
Live Loads, 3
Earthquake loads, 3, 173 emperical coefficients, 5
Elastomeric bearing pads, 154, 164 intensity reduction, 5
INDEX 1253

Loads Segmental box-girder bridge, 18, 83

dead, 3 anchorage blocks, 131
impact, 3 camber control, 195
wheel, 13 construction details, 141
Los Penasquitos Creek Bridge, 54 deflection, 197
erection, 88, 192
M flanges, 127
Mathivat, J., 164 flexural analysis, 85
Mission Valley Viaduct, 82, 172 hinges, 134
Moments live loads, 94
cantilever, 88, 96 piers, 135
secondary, 88, 93, 108 segment length, 132
redistribution due to creep, 13, 85 segment proportions, 126
unbalanced, 85 shear keys, 131
Muller, J., 98, 107 superstructure proportions, 119
Multi-beam bridge, 25 support details, 135
variable depth, I19
N variable width, 125
Napa River Bridge, 19, 1 23 webs, 127, 129
web stiffeners, 130
0 Segmental bridge, 15, 18, 83
Oldron Bridge, 123, 159, 190 Seismic forces, 173
Overhanging beams, 49 Shear flow, 62, 70
Shear lag, 67, 96
P Shear
Pacora River Bridge, 55 design, 145
Piers, 161 effect of variable depth, 148
flared, 171 prestressing for, 144
hollow prismatic, 163 Shear stress
solid, 162 closed section, 70
twin-wall, 168 open section, 69
Pine Valley Creek Bridge, 19, 122, 172
Shrinkage, 197
Poway Road Overcrossing, 53 French Code, 4
Prestressing material quantities, 203
Slab bridge, 24
Pucher, A., 9, 25, 46
Span length, 28
Spread box-beam bridge, 26
Q Strength analysis, 15 I
Quantities, prestressing materials, 203
allowable, 13
Redistribution of moment, 85, 96 tensile, 13
torsional, 64
Reinforcing, web, 102
shear, 69, 70, 145
RhGne-Alps Motorway, 194
Rio-Niteroi Bridge, 190 Substructure considerations, 161
Rombas Bridge, 194 Superstructure depth, 42, 57
Superstructure proportions
S box-girder bridge, 73
Saint Andre de Cubzac Bridge, 21, 125 girder bridge, 42
Saint Cloud Bridge, 21, 122 segmental box-girder bridge, 119
Secondary moments, 85, 88, 93, 108 Support constants, method of, 107, 217

T-beam, 16 Unbalanced moment. 85
Temperature, 3, 241
Temperature differential, 14, 102, 241
Tendon layout
box-girder bridge, 74, 76 V
segmental box-girder bridge, 104 Vail Pass, 121
Thenoz, M., 99 Volume changes, 4
Torsional constant, 62
Torsional stiffness
box-girder bridges, 18, 57, 59
girder bridges, 29 W
segmental box-girder bridges, 84 West Mission Bay Drive Bridge, 55
Transverse flexure Westergaard, H. M., 7
box-girder bridge, 59 Wide superstructures, 96
girder bridges, 30 box-girder bridges, 67
segmental box-girder bridges, 100 girder bridges, 33
Travelers, 120 segmental box-girder bridges, 96
Traveling forms, 181 Wind load, 3
Twin tubular-girders, 105 Witecki, A. A., 150